From kwizart at gmail.com Wed Oct 1 15:46:53 2008 From: kwizart at gmail.com (KH KH) Date: Wed, 1 Oct 2008 17:46:53 +0200 Subject: [Fedora-livecd-list] French fr_FR LiveCD spins In-Reply-To: <485115AF.8090401@kanarip.com> References: <4850727E.4060506@gmail.com> <485115AF.8090401@kanarip.com> Message-ID: 2008/6/12 Jeroen van Meeuwen : > Nicolas Chauvet wrote: Sorry, I've not replied right after your answer but here are the > the timezone should already be applied by kickstart; I removed the cp in > %post, I'll submit them to the spin-kickstarts git repo later today. > >> The current problem with adding dictionaries and man-pages is >> that they are disabled in the included ks. >> (-man-pages-* etc ) >> So I expect it would be easier to have an "international - offical" ks >> that would allow us to add this kind of packages; without having theses >> packages removed first. > > There's fedora-livecd-desktop-default.ks for "international" purposes; it > includes @*-support, for example. Ok, I think it is fixed with fedora-livecd-desktop-default.ks So localized spin will not pick every language. But on the other side, fedora-livecd-desktop.ks which all localized .ks are supposed to base upon, is still preventing some packages to be selected: (-aspell-* -hunspell-* -man-pages-*) Thus, localized ks should link on %include fedora-live-base.ks and take the fedora-livecd-desktop-default.ks as a base for localization (like what I've done with this one: ) http://rpms.kwizart.net/fedora/9/ISO/ks/livecd-fedora-9-desktop-fr_FR-20080912.ks Is it a better base than the one attached ? >> The %__install_lang_macro seems really interesting for localization works. >> It will save around 200Mo from a liveCD/USB spin. and will make the KDE Live >> spin ... a "potential" LiveCD spin. >> > > It is! I'm thinking to add this to the default localized spin thing so that > these get less fat and people can include (large) dictionaries etc. It seems that the support for using this option has been committed recently. Here is what I was using previously: --------------- %pre # Only install the needed locales - disabled mkdir -p /etc/rpm echo "%_install_langs en:fr:fr_FR:fr-FR" > /etc/rpm/macros.lang # Don't install docs - disabled echo "%_excludedocs 1" > /etc/rpm/macros.nodoc %end --------------- Any idea on what would be the replacement command to support this within a ks? Also, using %_excludedocs will conflict with the cp $INSTALL_ROOT/usr/share/doc/HTML/readme-live-image/fr_FR/readme-live-image-fr_FR.txt $LIVE_ROOT/README Maybe it would be better to have localized README not stored as %doc in fedora-release-notes ? Nicolas (kwizart) ps : I'm still testing about the echo fr-latin9 > /etc/X11/Xkbmap tweak i'm using in %post. But keyboard and date aren't set right on the international Fedora-10-beta-desktop livecd. That was the same problem with Fedora-9 gold. Now this tweak seems better than to hack via xorg.conf which we aren't supposed to use anyway. (KDE Lives may use another system). > Kind regards, > > Jeroen van Meeuwen > -------------- next part -------------- A non-text attachment was scrubbed... Name: fedora-livecd-desktop-fr_FR.ks Type: application/octet-stream Size: 1011 bytes Desc: not available URL: From todd_cooper at yahoo.com Wed Oct 1 16:13:50 2008 From: todd_cooper at yahoo.com (Todd Cooper) Date: Wed, 1 Oct 2008 09:13:50 -0700 (PDT) Subject: [Fedora-livecd-list] xconfig does not work Message-ID: <146335.47147.qm@web62407.mail.re1.yahoo.com> I added small things to the desktop kickstart (attached) and the xconfig no longer works. The biggest change is adding mysql to the rpms to be added. The .ks file attached has xconfig commented out. When I boot it comes up with the X login and when I click on the login user, it starts the login and then goes back to the login screen. Even if I login as root it does the same thing. Help? Thanks. -Todd Cooper -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: base-mysql.ks Type: application/octet-stream Size: 6310 bytes Desc: not available URL: From sundaram at fedoraproject.org Wed Oct 1 17:11:00 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Wed, 01 Oct 2008 22:41:00 +0530 Subject: [Fedora-livecd-list] Generic user and scripts Message-ID: <48E3AF24.2090400@fedoraproject.org> Hi, The Fedora kickstart files setup a user called "Fedora" and runs some init scripts under the same name. In my remix, I have retained the same name but it might be beneficial for such remixes and any other derivatives if we used something generic such as "system" instead of "Fedora". Comments? Rahul From aroldoamorim at ymail.com Thu Oct 2 02:28:05 2008 From: aroldoamorim at ymail.com (aroldo amorim) Date: Wed, 1 Oct 2008 19:28:05 -0700 (PDT) Subject: [Fedora-livecd-list] fedora-livecd Message-ID: <982544.80188.qm@web59714.mail.ac4.yahoo.com> hi request fron live CD tanks Novos endere?os, o Yahoo! que voc? conhece. Crie um email novo com a sua cara @ymail.com ou @rocketmail.com. http://br.new.mail.yahoo.com/addresses -------------- next part -------------- An HTML attachment was scrubbed... URL: From katzj at redhat.com Thu Oct 2 14:38:33 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 02 Oct 2008 10:38:33 -0400 Subject: [Fedora-livecd-list] Generic user and scripts In-Reply-To: <48E3AF24.2090400@fedoraproject.org> References: <48E3AF24.2090400@fedoraproject.org> Message-ID: <1222958313.16255.10.camel@aglarond.local> On Wed, 2008-10-01 at 22:41 +0530, Rahul Sundaram wrote: > The Fedora kickstart files setup a user called "Fedora" and runs some > init scripts under the same name. In my remix, I have retained the same > name but it might be beneficial for such remixes and any other > derivatives if we used something generic such as "system" instead of > "Fedora". Comments? If it's changed, it should probably be to something like 'liveuser' -- 'system' has a pretty distinct meaning when referring to users Jeremy From katzj at redhat.com Thu Oct 2 14:39:35 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 02 Oct 2008 10:39:35 -0400 Subject: [Fedora-livecd-list] xconfig does not work In-Reply-To: <146335.47147.qm@web62407.mail.re1.yahoo.com> References: <146335.47147.qm@web62407.mail.re1.yahoo.com> Message-ID: <1222958375.16255.11.camel@aglarond.local> On Wed, 2008-10-01 at 09:13 -0700, Todd Cooper wrote: > I added small things to the desktop kickstart (attached) and the > xconfig no longer works. > The biggest change is adding mysql to the rpms to be added. > The .ks file attached has xconfig commented out. > > When I boot it comes up with the X login and when I click on the login > user, it starts the login and then goes back to the login screen. > Even if I login as root it does the same thing. You probably are missing whatever the xsession is trying to run. If you switch to a tty and check ~/.xsession-errors, you'll probably see something more enlightening Jeremy From sundaram at fedoraproject.org Thu Oct 2 14:45:50 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Thu, 02 Oct 2008 20:15:50 +0530 Subject: [Fedora-livecd-list] Generic user and scripts In-Reply-To: <1222958313.16255.10.camel@aglarond.local> References: <48E3AF24.2090400@fedoraproject.org> <1222958313.16255.10.camel@aglarond.local> Message-ID: <48E4DE9E.3080703@fedoraproject.org> Jeremy Katz wrote: > On Wed, 2008-10-01 at 22:41 +0530, Rahul Sundaram wrote: >> The Fedora kickstart files setup a user called "Fedora" and runs some >> init scripts under the same name. In my remix, I have retained the same >> name but it might be beneficial for such remixes and any other >> derivatives if we used something generic such as "system" instead of >> "Fedora". Comments? > > If it's changed, it should probably be to something like 'liveuser' -- > 'system' has a pretty distinct meaning when referring to users Agreed. Would you do it? Rahul From katzj at redhat.com Thu Oct 2 15:08:41 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 02 Oct 2008 11:08:41 -0400 Subject: [Fedora-livecd-list] Generic user and scripts In-Reply-To: <48E4DE9E.3080703@fedoraproject.org> References: <48E3AF24.2090400@fedoraproject.org> <1222958313.16255.10.camel@aglarond.local> <48E4DE9E.3080703@fedoraproject.org> Message-ID: <1222960121.16255.15.camel@aglarond.local> On Thu, 2008-10-02 at 20:15 +0530, Rahul Sundaram wrote: > Jeremy Katz wrote: > > On Wed, 2008-10-01 at 22:41 +0530, Rahul Sundaram wrote: > >> The Fedora kickstart files setup a user called "Fedora" and runs some > >> init scripts under the same name. In my remix, I have retained the same > >> name but it might be beneficial for such remixes and any other > >> derivatives if we used something generic such as "system" instead of > >> "Fedora". Comments? > > > > If it's changed, it should probably be to something like 'liveuser' -- > > 'system' has a pretty distinct meaning when referring to users > > Agreed. Would you do it? Sure, done and pushed. I tried to make sure I got all of the references but it's possible I missed something Jeremy From kanarip at kanarip.com Thu Oct 2 15:16:40 2008 From: kanarip at kanarip.com (Jeroen van Meeuwen) Date: Thu, 02 Oct 2008 17:16:40 +0200 Subject: [Fedora-livecd-list] Generic user and scripts In-Reply-To: <1222960121.16255.15.camel@aglarond.local> References: <48E3AF24.2090400@fedoraproject.org> <1222958313.16255.10.camel@aglarond.local> <48E4DE9E.3080703@fedoraproject.org> <1222960121.16255.15.camel@aglarond.local> Message-ID: <48E4E5D8.8080709@kanarip.com> Jeremy Katz wrote: > On Thu, 2008-10-02 at 20:15 +0530, Rahul Sundaram wrote: >> Jeremy Katz wrote: >>> On Wed, 2008-10-01 at 22:41 +0530, Rahul Sundaram wrote: >>>> The Fedora kickstart files setup a user called "Fedora" and runs some >>>> init scripts under the same name. In my remix, I have retained the same >>>> name but it might be beneficial for such remixes and any other >>>> derivatives if we used something generic such as "system" instead of >>>> "Fedora". Comments? >>> If it's changed, it should probably be to something like 'liveuser' -- >>> 'system' has a pretty distinct meaning when referring to users >> Agreed. Would you do it? > > Sure, done and pushed. I tried to make sure I got all of the references > but it's possible I missed something > For those of you who want to watch the commits to the spin-kickstarts GIT repository; https://fedorahosted.org/mailman/listinfo/spin-kickstarts-commits I thought I'd just mention it. Kind regards, Jeroen van Meeuwen -kanarip From justin at grangerx.com Thu Oct 2 15:36:47 2008 From: justin at grangerx.com (Justin Hochstetler) Date: Thu, 2 Oct 2008 11:36:47 -0400 Subject: [Fedora-livecd-list] Persistent Overlay on LiveCD Message-ID: <2575ced0810020836w47e10171uc982a3fcd7a5573f@mail.gmail.com> Hi all, Apologies if this is in a FAQ somewhere; I didn't see it if so. Is there a way with Fedora 9 or Fedora 10 to use a LiveCD and have a persistent overlay on another device, such as (obviously) a USB stick, without also installing Fedora itself to the stick? Some of the PCs I work with boot the CD okay, but tend to not like the USB as much, at least where booting-from is concerned. Thanks, Justin From pedrofsilva at netcabo.pt Fri Oct 3 01:47:58 2008 From: pedrofsilva at netcabo.pt (Pedro Silva) Date: Fri, 03 Oct 2008 02:47:58 +0100 Subject: [Fedora-livecd-list] Xorg keyboard patch Message-ID: <1222998478.3390.8.camel@localhost.localdomain> Hi! Hinted by Evdev and new gdm options, I tested gnome based kickstart files from git without the xorg keyboard patch and it works. I can sucessfully choose a portuguese keyboard in gdm and after entering in a gnome session, applications correctly detect my keyboard. Also, gdm is presented in portuguese language. I committed these changes to pt_PT kickstart files based on gnome. I also fixed the template/patch for pt_PT language. Hopefully, I will have the same results on all pt_PT ks files and the patch is no longer needed. GDM looks great with those new features. :) Best regards, Pedro Silva -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: Esta ? uma parte de mensagem assinada digitalmente URL: From katzj at fedoraproject.org Fri Oct 3 02:28:03 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Fri, 3 Oct 2008 02:28:03 +0000 (UTC) Subject: [Fedora-livecd-list] imgcreate/live.py Message-ID: <20081003022803.69B16C07B7@lists.fedorahosted.org> imgcreate/live.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) New commits: commit cb364b5faeecb22bc8ecd73dd94feee4cf3ad4e6 Author: Jeremy Katz Date: Tue Sep 30 12:15:42 2008 -0400 SD needs sdhci-pci now as well diff --git a/imgcreate/live.py b/imgcreate/live.py index f98f76c..17d3037 100644 --- a/imgcreate/live.py +++ b/imgcreate/live.py @@ -220,7 +220,7 @@ class LiveImageCreatorBase(LoopImageCreator): f.write('MODULES+="firewire-sbp2 firewire-ohci "\n') f.write('MODULES+="sbp2 ohci1394 ieee1394 "\n') elif module == "=mmc": - f.write('MODULES+="mmc_block sdhci "\n') + f.write('MODULES+="mmc_block sdhci sdhci-pci "\n') elif module == "=pcmcia": f.write('MODULES+="pata_pcmcia "\n') else: From katzj at fedoraproject.org Fri Oct 3 02:29:25 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Fri, 3 Oct 2008 02:29:25 +0000 (UTC) Subject: [Fedora-livecd-list] 3 commits - tools/livecd-iso-to-disk.sh Message-ID: <20081003022925.31952C07B7@lists.fedorahosted.org> tools/livecd-iso-to-disk.sh | 36 ++++++++++++++++++++++++++++++------ 1 file changed, 30 insertions(+), 6 deletions(-) New commits: commit 79ad977f7ae66dd4cdcbc6e3e6c224711a65eca1 Author: Jeremy Katz Date: Thu Oct 2 22:24:47 2008 -0400 Set up persistent /home for the XO on the internal nand Unless explicitly requested otherwise with --xo-no-home, use the /home on the XO's NAND by default for /home. This helps us to keep the memory footprint a little bit lower diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index 27e59ec..5c20640 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -266,6 +266,9 @@ while [ $# -gt 2 ]; do xo=1 xousb=1 ;; + --xo-no-home) + xonohome=1 + ;; --extra-kernel-args) kernelargs=$2 shift @@ -496,6 +499,9 @@ fi if [ -n "$xo" ]; then echo "Setting up /olpc-usb.fth file" args=$(egrep "^[ ]*append" $USBMNT/$SYSLINUXPATH/isolinux.cfg |head -n1 |sed -e 's/.*initrd=[^ ]*//') + if [ -n "$xonohome" -a ! -f $USBMNT/LiveOS/$HOMEFILE ]; then + args="$args persistenthome=mtd0" + fi cat > $USBMNT/olpc-usb.fth < Date: Thu Oct 2 22:24:05 2008 -0400 Switch all references to home.img to be $HOMEFILE diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index 7cb97af..27e59ec 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -226,6 +226,7 @@ fi cryptedhome=1 keephome=1 +HOMEFILE="home.img" while [ $# -gt 2 ]; do case $1 in --overlay-size-mb) @@ -341,7 +342,7 @@ mount $mountopts $USBDEV $USBMNT || exitclean trap exitclean SIGINT SIGTERM -if [ -f "$USBMNT/LiveOS/home.img" -a -n "$keephome" -a -n "$homesizemb" ]; then +if [ -f "$USBMNT/LiveOS/$HOMEFILE" -a -n "$keephome" -a -n "$homesizemb" ]; then echo "ERROR: Requested keeping existing /home and specified a size for /home" echo "Please either don't specify a size or specify --delete-home" exitclean @@ -355,7 +356,7 @@ else fi if [ -d $USBMNT/LiveOS ]; then tbd=$(du -s -B 1M $USBMNT/LiveOS | awk {'print $1;'}) - [ -f $USBMNT/LiveOS/home.img ] && homesz=$(du -s -B 1M $USBMNT/LiveOS/home.img | awk {'print $1;'}) + [ -f $USBMNT/LiveOS/$HOMEFILE ] && homesz=$(du -s -B 1M $USBMNT/LiveOS/$HOMEFILE | awk {'print $1;'}) [ -n "$homesz" -a -n "$keephome" ] && tbd=$(($tbd - $homesz)) else tbd=0 @@ -374,7 +375,7 @@ fi if [ -d $USBMNT/LiveOS -a -z "$force" ]; then echo "Already set up as live image." - if [ -z "$keephome" -a -e $USBMNT/LiveOS/home.img ]; then + if [ -z "$keephome" -a -e $USBMNT/LiveOS/$HOMEFILE ]; then echo "WARNING: Persistent /home will be deleted!!!" echo "Press Enter to continue or ctrl-c to abort" read @@ -382,7 +383,7 @@ if [ -d $USBMNT/LiveOS -a -z "$force" ]; then echo "Deleting old OS in fifteen seconds..." sleep 15 - [ -e "$USBMNT/LiveOS/home.img" -a -n "$keephome" ] && mv $USBMNT/LiveOS/home.img $USBMNT/home.img + [ -e "$USBMNT/LiveOS/$HOMEFILE" -a -n "$keephome" ] && mv $USBMNT/LiveOS/$HOMEFILE $USBMNT/$HOMEFILE fi rm -rf $USBMNT/LiveOS @@ -392,7 +393,7 @@ echo "Copying live image to USB stick" [ -z "$mactel" -a ! -d $USBMNT/$SYSLINUXPATH ] && mkdir -p $USBMNT/$SYSLINUXPATH [ -n "$mactel" -a ! -d $USBMNT/EFI/boot ] && mkdir -p $USBMNT/EFI/boot [ ! -d $USBMNT/LiveOS ] && mkdir $USBMNT/LiveOS -[ -n "$keephome" -a -f "$USBMNT/home.img" ] && mv $USBMNT/home.img $USBMNT/LiveOS/home.img +[ -n "$keephome" -a -f "$USBMNT/$HOMEFILE" ] && mv $USBMNT/$HOMEFILE $USBMNT/LiveOS/$HOMEFILE # cases without /LiveOS are legacy detection, remove for F10 if [ -f $CDMNT/LiveOS/squashfs.img ]; then cp $CDMNT/LiveOS/squashfs.img $USBMNT/LiveOS/squashfs.img || exitclean @@ -465,7 +466,6 @@ fi if [ -n "$homesizemb" ]; then echo "Initializing persistent /home" - HOMEFILE=home.img if [ "$USBFS" = "vfat" ]; then # vfat can't handle sparse files dd if=/dev/zero of=$USBMNT/LiveOS/$HOMEFILE count=$homesizemb bs=1M commit e4b03a3e249ecf550b40b659914b84210989b0e3 Author: Jeremy Katz Date: Tue Sep 23 13:01:51 2008 -0400 Temporarily patch up ramdisk handling for XO via olpc.fth (from Mitch Bradley) The XO firmware is loading the initrd at the wrong address which is what was breaking the use of standard Fedora kernels. This does runtime patching of ofw from the boot file until there's a new firmware revision that we can require. OFW code is from Mitch Bradley diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index b9b877b..7cb97af 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -498,6 +498,15 @@ if [ -n "$xo" ]; then args=$(egrep "^[ ]*append" $USBMNT/$SYSLINUXPATH/isolinux.cfg |head -n1 |sed -e 's/.*initrd=[^ ]*//') cat > $USBMNT/olpc-usb.fth < $USBMNT/olpc-sd.fth < How do I switch to a tty from the GUI? Is there a key sequence for X windows once booted to go back to a regular login prompt? Todd M. Cooper todd_cooper at yahoo.com ----- Original Message ---- From: Jeremy Katz To: Todd Cooper ; fedora-livecd-list at redhat.com Sent: Thursday, October 2, 2008 10:39:35 AM Subject: Re: [Fedora-livecd-list] xconfig does not work On Wed, 2008-10-01 at 09:13 -0700, Todd Cooper wrote: > I added small things to the desktop kickstart (attached) and the > xconfig no longer works. > The biggest change is adding mysql to the rpms to be added. > The .ks file attached has xconfig commented out. > > When I boot it comes up with the X login and when I click on the login > user, it starts the login and then goes back to the login screen. > Even if I login as root it does the same thing. You probably are missing whatever the xsession is trying to run. If you switch to a tty and check ~/.xsession-errors, you'll probably see something more enlightening Jeremy -- Fedora-livecd-list mailing list Fedora-livecd-list at redhat.com https://www.redhat.com/mailman/listinfo/fedora-livecd-list -------------- next part -------------- An HTML attachment was scrubbed... URL: From gary at mlbassoc.com Fri Oct 3 14:53:21 2008 From: gary at mlbassoc.com (Gary Thomas) Date: Fri, 03 Oct 2008 08:53:21 -0600 Subject: [Fedora-livecd-list] xconfig does not work In-Reply-To: <84782.93050.qm@web62405.mail.re1.yahoo.com> References: <84782.93050.qm@web62405.mail.re1.yahoo.com> Message-ID: <48E631E1.6000701@mlbassoc.com> Todd Cooper wrote: > How do I switch to a tty from the GUI? > Is there a key sequence for X windows once booted to go back to a > regular login prompt? > Ctl-Atl-F8, followed by Alt-F1..Alt-Fn to switch to virtual terminal 'n'. Ctl-Alt-F7 gets you back to X > ----- Original Message ---- > From: Jeremy Katz > To: Todd Cooper ; fedora-livecd-list at redhat.com > Sent: Thursday, October 2, 2008 10:39:35 AM > Subject: Re: [Fedora-livecd-list] xconfig does not work > > On Wed, 2008-10-01 at 09:13 -0700, Todd Cooper wrote: >> I added small things to the desktop kickstart (attached) and the >> xconfig no longer works. >> The biggest change is adding mysql to the rpms to be added. >> The .ks file attached has xconfig commented out. >> >> When I boot it comes up with the X login and when I click on the login >> user, it starts the login and then goes back to the login screen. >> Even if I login as root it does the same thing. > > You probably are missing whatever the xsession is trying to run. If you > switch to a tty and check ~/.xsession-errors, you'll probably see > something more enlightening -- ------------------------------------------------------------ Gary Thomas | Consulting for the MLB Associates | Embedded world ------------------------------------------------------------ From sundaram at fedoraproject.org Fri Oct 3 21:08:51 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Sat, 04 Oct 2008 02:38:51 +0530 Subject: [Fedora-livecd-list] Livecd-creator and source rpms Message-ID: <48E689E3.6030400@fedoraproject.org> Hi, If I am building and distributing a ISO containing Fedora packages, I would also want to distribute the equivalent source ISO. Unlike Pungi which does this by default, livecd-creator doesn't seem to have a easy way to do this. It especially becomes tricky if I am including not just the GA packages but also including updates. Can livecd-creator be modified to build source ISO images as well? Rahul From jkeating at redhat.com Fri Oct 3 21:20:28 2008 From: jkeating at redhat.com (Jesse Keating) Date: Fri, 03 Oct 2008 14:20:28 -0700 Subject: [Fedora-livecd-list] Livecd-creator and source rpms In-Reply-To: <48E689E3.6030400@fedoraproject.org> References: <48E689E3.6030400@fedoraproject.org> Message-ID: <1223068828.9987.5.camel@luminos.localdomain> On Sat, 2008-10-04 at 02:38 +0530, Rahul Sundaram wrote: > If I am building and distributing a ISO containing Fedora packages, I > would also want to distribute the equivalent source ISO. Unlike Pungi > which does this by default, livecd-creator doesn't seem to have a easy > way to do this. It especially becomes tricky if I am including not just > the GA packages but also including updates. Can livecd-creator be > modified to build source ISO images as well? I'm sure it could. Somebody just has to do the work to add that functionality. I'm sure some of the pungi code could be stolen or made a generic utility and used directly to determine which srpms to grab. -- Jesse Keating Fedora -- Freedom? is a feature! identi.ca: http://identi.ca/jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part URL: From sundaram at fedoraproject.org Fri Oct 3 21:22:46 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Sat, 04 Oct 2008 02:52:46 +0530 Subject: [Fedora-livecd-list] Livecd-creator and source rpms In-Reply-To: <1223068828.9987.5.camel@luminos.localdomain> References: <48E689E3.6030400@fedoraproject.org> <1223068828.9987.5.camel@luminos.localdomain> Message-ID: <48E68D26.7060404@fedoraproject.org> Jesse Keating wrote: > On Sat, 2008-10-04 at 02:38 +0530, Rahul Sundaram wrote: >> If I am building and distributing a ISO containing Fedora packages, I >> would also want to distribute the equivalent source ISO. Unlike Pungi >> which does this by default, livecd-creator doesn't seem to have a easy >> way to do this. It especially becomes tricky if I am including not just >> the GA packages but also including updates. Can livecd-creator be >> modified to build source ISO images as well? > > I'm sure it could. Somebody just has to do the work to add that > functionality. I'm sure some of the pungi code could be stolen or made > a generic utility and used directly to determine which srpms to grab. Could the common code be split out or IIRC, there was some mumbling about merging pungi and livecd-creator. What are the current thoughts on that? Rahul From bkearney at redhat.com Fri Oct 3 21:24:12 2008 From: bkearney at redhat.com (Bryan Kearney) Date: Fri, 03 Oct 2008 17:24:12 -0400 Subject: [Fedora-livecd-list] Livecd-creator and source rpms In-Reply-To: <48E68D26.7060404@fedoraproject.org> References: <48E689E3.6030400@fedoraproject.org> <1223068828.9987.5.camel@luminos.localdomain> <48E68D26.7060404@fedoraproject.org> Message-ID: <48E68D7C.6060006@redhat.com> Rahul Sundaram wrote: > Jesse Keating wrote: >> On Sat, 2008-10-04 at 02:38 +0530, Rahul Sundaram wrote: >>> If I am building and distributing a ISO containing Fedora packages, I >>> would also want to distribute the equivalent source ISO. Unlike Pungi >>> which does this by default, livecd-creator doesn't seem to have a >>> easy way to do this. It especially becomes tricky if I am including >>> not just the GA packages but also including updates. Can >>> livecd-creator be modified to build source ISO images as well? >> >> I'm sure it could. Somebody just has to do the work to add that >> functionality. I'm sure some of the pungi code could be stolen or made >> a generic utility and used directly to determine which srpms to grab. > > Could the common code be split out or IIRC, there was some mumbling > about merging pungi and livecd-creator. What are the current thoughts on > that? We had gotten the same requirement for the appliance building. We can get David Huff to help out. I will make sure he sees this thread :) -- bk From katzj at fedoraproject.org Fri Oct 3 23:57:44 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Fri, 3 Oct 2008 23:57:44 +0000 (UTC) Subject: [Fedora-livecd-list] tools/livecd-iso-to-disk.sh Message-ID: <20081003235744.3C703120195@lists.fedorahosted.org> tools/livecd-iso-to-disk.sh | 57 ++++++++++++++++---------------------------- 1 file changed, 22 insertions(+), 35 deletions(-) New commits: commit 9b779482c390b399cb4a8e7c656982a39b974461 Author: Jeremy Katz Date: Fri Oct 3 19:55:32 2008 -0400 Auto-determine USB vs SD on the XO Rather than requiring specifying whether you're doing USB or SD, we can see which the fth script is actually running from. Do some magic (again, thanks to Mitch Bradley) to do so diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index 5c20640..f385dcb 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -258,14 +258,6 @@ while [ $# -gt 2 ]; do --xo) xo=1 ;; - --xo-sd) - xo=1 - xosd=1 - ;; - --xo-usb) - xo=1 - xousb=1 - ;; --xo-no-home) xonohome=1 ;; @@ -497,12 +489,13 @@ fi # we'd do this unconditionally, but you have to have a kernel that will # boot on the XO anyway. if [ -n "$xo" ]; then - echo "Setting up /olpc-usb.fth file" + echo "Setting up /boot/olpc.fth file" args=$(egrep "^[ ]*append" $USBMNT/$SYSLINUXPATH/isolinux.cfg |head -n1 |sed -e 's/.*initrd=[^ ]*//') if [ -n "$xonohome" -a ! -f $USBMNT/LiveOS/$HOMEFILE ]; then args="$args persistenthome=mtd0" fi - cat > $USBMNT/olpc-usb.fth < $USBMNT/boot/olpc.fth < $USBMNT/olpc-sd.fth <= if + " sd:" + else + " u:" + then + " BOOTPATHDEV" \$set-macro ; -' high-ramdisk to load-ramdisk +set-bootpath-dev " $args" to boot-file -" sd:\syslinux\initrd0.img" to ramdisk +" \${BOOTPATHDEV}\syslinux\initrd0.img" expand$ to ramdisk +" \${BOOTPATHDEV}\syslinux\vmlinuz0" expand$ to boot-device unfreeze -boot sd:\syslinux\vmlinuz0 +boot EOF - if [ -n "$xosd" ]; then - mkdir $USBMNT/boot - cp $USBMNT/olpc-sd.fth $USBMNT/boot/olpc.fth - elif [ -n "$xousb" ]; then - mkdir $USBMNT/boot - cp $USBMNT/olpc-usb.fth $USBMNT/boot/olpc.fth - fi fi echo "Installing boot loader" From dex.mbox at googlemail.com Sat Oct 4 16:24:43 2008 From: dex.mbox at googlemail.com (dexter) Date: Sat, 4 Oct 2008 17:24:43 +0100 Subject: [Fedora-livecd-list] livecd-tools deps F9 Message-ID: Should livecd-creator Require libselinux-python in F9 ? ...dex From adam.stokes at gmail.com Sun Oct 5 17:09:22 2008 From: adam.stokes at gmail.com (Adam Stokes) Date: Sun, 5 Oct 2008 13:09:22 -0400 Subject: [Fedora-livecd-list] gdm autologin Message-ID: So I've setup /etc/gdm/custom.conf for autologin options as described here http://library.gnome.org/admin/gdm/stable/configuration.html.en#daemonconfig but with F9 gdm autologin only happens after clicking the desired user. Anyone else run into this problem or know of a better way to enable a sort of kiosk mode with F9 gdm? Thanks -- ( adam stokes ) || ( adam.stokes at gmail.com ) From igorsoares at gmail.com Sun Oct 5 21:50:37 2008 From: igorsoares at gmail.com (Igor Pires Soares) Date: Sun, 05 Oct 2008 18:50:37 -0300 Subject: [Fedora-livecd-list] Xorg keyboard patch In-Reply-To: <1222998478.3390.8.camel@localhost.localdomain> References: <1222998478.3390.8.camel@localhost.localdomain> Message-ID: <1223243437.3118.5.camel@amd5600> Em Sex, 2008-10-03 ?s 02:47 +0100, Pedro Silva escreveu: > Hi! > > Hinted by Evdev and new gdm options, I tested gnome based kickstart > files from git without the xorg keyboard patch and it works. I can > sucessfully choose a portuguese keyboard in gdm and after entering in a > gnome session, applications correctly detect my keyboard. Also, gdm is > presented in portuguese language. > > I committed these changes to pt_PT kickstart files based on gnome. I > also fixed the template/patch for pt_PT language. > > Hopefully, I will have the same results on all pt_PT ks files and the > patch is no longer needed. GDM looks great with those new features. :) Also works for me but GDM still shows an USA keyboard as default in the login screen, although the actual keyboard layout is set to br-abnt2. Regards, Igor Pires Soares From katzj at redhat.com Mon Oct 6 13:32:28 2008 From: katzj at redhat.com (Jeremy Katz) Date: Mon, 06 Oct 2008 09:32:28 -0400 Subject: [Fedora-livecd-list] gdm autologin In-Reply-To: References: Message-ID: <1223299948.19348.23.camel@aglarond.local> On Sun, 2008-10-05 at 13:09 -0400, Adam Stokes wrote: > So I've setup /etc/gdm/custom.conf for autologin options as described > here http://library.gnome.org/admin/gdm/stable/configuration.html.en#daemonconfig > but with F9 gdm autologin only happens after clicking the desired > user. Anyone else run into this problem or know of a better way to > enable a sort of kiosk mode with F9 gdm? I'm unsure if those docs are updated for the version of gdm being shipped in F9. But iirc, it should work if you do TimedLogin like we do on the existing live images and then also set TimedLoginDelay=0 Jeremy From jkeating at redhat.com Tue Oct 7 00:38:14 2008 From: jkeating at redhat.com (Jesse Keating) Date: Mon, 06 Oct 2008 17:38:14 -0700 Subject: [Fedora-livecd-list] Re-using cache between i386 and x86_64 composes Message-ID: <1223339894.9987.53.camel@luminos.localdomain> Maybe this is just a "me" thing and I should use different caches, but when I try to use the same cachedir and compose i686 and x86_64 images, I can run into a situation where the repomd.xml for say the 64bit repo is newer than the repomd.xml for the 32bit repo. This can lead to a situation where livecd-creator won't pull down the repodata for the 32bit repo after a 64bit run, which will lead to the compose crashing horribly. With pungi, I force re-getting all the repodata, regardless of existing content. It's a little extra downloading but it ensures we always get the right repodata. Do people think that's what should be done with livecd-creator, or should I just set up my scripts so that when composing i686 I use one cache dir, and when composing x86_64 I use another? -- Jesse Keating Fedora -- Freedom? is a feature! identi.ca: http://identi.ca/jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part URL: From katzj at redhat.com Tue Oct 7 00:49:12 2008 From: katzj at redhat.com (Jeremy Katz) Date: Mon, 06 Oct 2008 20:49:12 -0400 Subject: [Fedora-livecd-list] Re-using cache between i386 and x86_64 composes In-Reply-To: <1223339894.9987.53.camel@luminos.localdomain> References: <1223339894.9987.53.camel@luminos.localdomain> Message-ID: <1223340552.754.11.camel@aglarond.local> On Mon, 2008-10-06 at 17:38 -0700, Jesse Keating wrote: > Maybe this is just a "me" thing and I should use different caches, but > when I try to use the same cachedir and compose i686 and x86_64 images, > I can run into a situation where the repomd.xml for say the 64bit repo > is newer than the repomd.xml for the 32bit repo. This can lead to a > situation where livecd-creator won't pull down the repodata for the > 32bit repo after a 64bit run, which will lead to the compose crashing > horribly. > > With pungi, I force re-getting all the repodata, regardless of existing > content. It's a little extra downloading but it ensures we always get > the right repodata. Do people think that's what should be done with > livecd-creator, or should I just set up my scripts so that when > composing i686 I use one cache dir, and when composing x86_64 I use > another? The default is that we use a fresh cache directory for each individual run. If you're reusing cachedirs, you should either be keeping them arch-specific or just nuke metadata yourself in between runs. I'd really prefer keeping any information about how repos are built, verified, etc out of livecd-creator as it's just another thing that can get out of sync over time Jeremy From jkeating at j2solutions.net Tue Oct 7 00:59:59 2008 From: jkeating at j2solutions.net (Jesse Keating) Date: Mon, 06 Oct 2008 17:59:59 -0700 Subject: [Fedora-livecd-list] Re-using cache between i386 and x86_64 composes In-Reply-To: <1223340552.754.11.camel@aglarond.local> References: <1223339894.9987.53.camel@luminos.localdomain> <1223340552.754.11.camel@aglarond.local> Message-ID: <1223341199.9987.62.camel@luminos.localdomain> On Mon, 2008-10-06 at 20:49 -0400, Jeremy Katz wrote: > The default is that we use a fresh cache directory for each individual > run. If you're reusing cachedirs, you should either be keeping them > arch-specific or just nuke metadata yourself in between runs. I'd > really prefer keeping any information about how repos are built, > verified, etc out of livecd-creator as it's just another thing that can > get out of sync over time > I wouldn't think that unconditionally cleaning out repodata (at least the repomd.xml file) each run would be 'information about how repos are built, verified, etc' but maybe that depends on the 'etc'. -- Jesse Keating RHCE (http://jkeating.livejournal.com) Fedora Project (http://fedoraproject.org/wiki/JesseKeating) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) identi.ca (http://identi.ca/jkeating) -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part URL: From katzj at redhat.com Tue Oct 7 01:21:32 2008 From: katzj at redhat.com (Jeremy Katz) Date: Mon, 06 Oct 2008 21:21:32 -0400 Subject: [Fedora-livecd-list] Re-using cache between i386 and x86_64 composes In-Reply-To: <1223341199.9987.62.camel@luminos.localdomain> References: <1223339894.9987.53.camel@luminos.localdomain> <1223340552.754.11.camel@aglarond.local> <1223341199.9987.62.camel@luminos.localdomain> Message-ID: <1223342492.754.14.camel@aglarond.local> On Mon, 2008-10-06 at 17:59 -0700, Jesse Keating wrote: > On Mon, 2008-10-06 at 20:49 -0400, Jeremy Katz wrote: > > The default is that we use a fresh cache directory for each individual > > run. If you're reusing cachedirs, you should either be keeping them > > arch-specific or just nuke metadata yourself in between runs. I'd > > really prefer keeping any information about how repos are built, > > verified, etc out of livecd-creator as it's just another thing that can > > get out of sync over time > > > I wouldn't think that unconditionally cleaning out repodata (at least > the repomd.xml file) each run would be 'information about how repos are > built, verified, etc' but maybe that depends on the 'etc'. The fact that it's repomd.xml is an implementation-detail that there's no real business for livecd-creator to need to know. And while repomd.xml hasn't changed, we definitely have changed similar things in the past (rest of the metadata being xml -> sqlite) Jeremy From jkeating at redhat.com Tue Oct 7 01:32:02 2008 From: jkeating at redhat.com (Jesse Keating) Date: Mon, 06 Oct 2008 18:32:02 -0700 Subject: [Fedora-livecd-list] Re-using cache between i386 and x86_64 composes In-Reply-To: <1223342492.754.14.camel@aglarond.local> References: <1223339894.9987.53.camel@luminos.localdomain> <1223340552.754.11.camel@aglarond.local> <1223341199.9987.62.camel@luminos.localdomain> <1223342492.754.14.camel@aglarond.local> Message-ID: <1223343122.9987.66.camel@luminos.localdomain> On Mon, 2008-10-06 at 21:21 -0400, Jeremy Katz wrote: > The fact that it's repomd.xml is an implementation-detail that there's > no real business for livecd-creator to need to know. And while > repomd.xml hasn't changed, we definitely have changed similar things in > the past (rest of the metadata being xml -> sqlite) Er, I'm not asking for a manual rm, rather a yum API call to clean out the repodata, surely that's not too specific? -- Jesse Keating Fedora -- Freedom? is a feature! identi.ca: http://identi.ca/jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part URL: From jkeating at redhat.com Tue Oct 7 17:51:14 2008 From: jkeating at redhat.com (Jesse Keating) Date: Tue, 07 Oct 2008 10:51:14 -0700 Subject: [Fedora-livecd-list] @group --something support is broken Message-ID: <1223401874.4410.16.camel@luminos.localdomain> Looking at the code, I don't think this ever worked quite right, and there is a bit of API confusion. From yuminst.py: def selectGroup(self, grp, include = pykickstart.parser.GROUP_DEFAULT): yum.YumBase.selectGroup(self, grp) if include == pykickstart.parser.GROUP_REQUIRED: map(lambda p: self.deselectPackage(p), grp.default_packages.keys()) elif include == pykickstart.parser.GROUP_ALL: map(lambda p: self.selectPackage(p), grp.optional_packages.keys()) There is a bit of a problem in this function. 'grp' right now is being passed in (via creator.py) as a string that is the group name, and that works great for the first line of the function, where yum expects a string for selectGroup. However if there were options passed to the group and we have to fetch values, we try to call 'grp.default_packages.keys()' or optional_packages. A string doesn't have those attributes, but a pykickstart group object does. Fixing this is easy, just make 'grp' the actual pykickstart group object. In fact, if we did that, we could do away with the second argument, include =. That would be the most simple fix, but that would break API. Even just changing 'grp' to be expected to be a pykickstart group object could be seen as an API break. I don't really see a way to fix this without some sort of API change, unless we add yet another keyword argument to the function that is the pykickstart object, and declare that unless you pass that third option, you don't get --foo support for your groups. Not that it ever worked anyway. Other options are create a new function all together, which adds to the API, and callers that want groups with options can use the new call, or we could try to detect what object 'grp' is within the existing function and operate accordingly (although I don't see how you can pass in a string and still make --foo options work, at least not without a lot of extra yum object calls to get the group and get the optional/all members). Since this is a sticky fix, I thought I'd post to the list for discussion before working on a patch. Thoughts? -- Jesse Keating Fedora -- Freedom? is a feature! identi.ca: http://identi.ca/jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part URL: From katzj at redhat.com Tue Oct 7 18:39:05 2008 From: katzj at redhat.com (Jeremy Katz) Date: Tue, 07 Oct 2008 14:39:05 -0400 Subject: [Fedora-livecd-list] @group --something support is broken In-Reply-To: <1223401874.4410.16.camel@luminos.localdomain> References: <1223401874.4410.16.camel@luminos.localdomain> Message-ID: <1223404745.754.36.camel@aglarond.local> On Tue, 2008-10-07 at 10:51 -0700, Jesse Keating wrote: > Looking at the code, I don't think this ever worked quite right, and > there is a bit of API confusion. It hasn't been touched since the initial API work was done, so it's entirely likely it's just something that got broken with all the code churn that happened there > Fixing this is easy, just make 'grp' the actual pykickstart group > object. In fact, if we did that, we could do away with the second > argument, include =. That would be the most simple fix, but that would > break API. The pykickstart group doesn't have the required_packages bits. Instead, you'd need the yum group. Given what callers are doing as well as what the yum API call expects, I think that a string is what it was always intended to be and we just lost something to grab the yum group. Although looking at current yum API, we could just go off of the include constant and set group_package_types to the appropriate subset of (mandatory, default, optional) and make things even a little clearer. Jeremy From jkeating at redhat.com Tue Oct 7 23:42:44 2008 From: jkeating at redhat.com (Jesse Keating) Date: Tue, 7 Oct 2008 16:42:44 -0700 Subject: [Fedora-livecd-list] [PATCH] Fix a bug with getting groups with options (@foo --optional). In-Reply-To: <1223404745.754.36.camel@aglarond.local> References: <1223404745.754.36.camel@aglarond.local> Message-ID: <1223422964-1096-1-git-send-email-jkeating@redhat.com> New yum api (in 3.2.x) allows us to pass into selectGroups what set we want from the group. The default is mandatory and default. Use that unless we have options from kickstart for either optional packages (IE everything in the group) or just the default packages of the group. This also allows us to just get the right packages the first time rather than getting packages and then pruning/adding as necessary. Less trips through the yum stack, which is a good thing. --- imgcreate/yuminst.py | 9 ++++++--- 1 files changed, 6 insertions(+), 3 deletions(-) diff --git a/imgcreate/yuminst.py b/imgcreate/yuminst.py index a7b04b6..b89a6ed 100644 --- a/imgcreate/yuminst.py +++ b/imgcreate/yuminst.py @@ -110,11 +110,14 @@ class LiveCDYum(yum.YumBase): logging.warn("No such package %s to remove" %(pkg,)) def selectGroup(self, grp, include = pykickstart.parser.GROUP_DEFAULT): - yum.YumBase.selectGroup(self, grp) + # default to getting mandatory and default packages from a group + # unless we have specific options from kickstart + package_types = ['mandatory', 'default'] if include == pykickstart.parser.GROUP_REQUIRED: - map(lambda p: self.deselectPackage(p), grp.default_packages.keys()) + package_types.remove('default') elif include == pykickstart.parser.GROUP_ALL: - map(lambda p: self.selectPackage(p), grp.optional_packages.keys()) + package_types.append('optional') + yum.YumBase.selectGroup(self, grp, group_package_types=package_types) def addRepository(self, name, url = None, mirrorlist = None): def _varSubstitute(option): -- 1.6.0.1 From jkeating at redhat.com Tue Oct 7 23:47:18 2008 From: jkeating at redhat.com (Jesse Keating) Date: Tue, 07 Oct 2008 16:47:18 -0700 Subject: [Fedora-livecd-list] [PATCH] Fix a bug with getting groups with options (@foo --optional). In-Reply-To: <1223422964-1096-1-git-send-email-jkeating@redhat.com> References: <1223404745.754.36.camel@aglarond.local> <1223422964-1096-1-git-send-email-jkeating@redhat.com> Message-ID: <1223423238.4410.39.camel@luminos.localdomain> On Tue, 2008-10-07 at 16:42 -0700, Jesse Keating wrote: > New yum api (in 3.2.x) Hrm, taking this in will bump up our minimum yum level listed in the .spec file, but 3.2 is in use back to Fedora 8, so I don't think that would be too much of a deal. -- Jesse Keating Fedora -- Freedom? is a feature! identi.ca: http://identi.ca/jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part URL: From pedrofsilva at netcabo.pt Wed Oct 8 12:26:35 2008 From: pedrofsilva at netcabo.pt (Pedro Silva) Date: Wed, 08 Oct 2008 13:26:35 +0100 Subject: [Fedora-livecd-list] Xorg keyboard patch In-Reply-To: <1223243437.3118.5.camel@amd5600> References: <1222998478.3390.8.camel@localhost.localdomain> <1223243437.3118.5.camel@amd5600> Message-ID: <1223468795.3088.4.camel@localhost.localdomain> Dom, 2008-10-05 ?s 18:50 -0300, Igor Pires Soares escreveu: > Also works for me but GDM still shows an USA keyboard as default in the > login screen, although the actual keyboard layout is set to br-abnt2. I googled trying to find some option I could put in /etc/gdm.conf but couldn't find anything. Anyone, is it possible to define a default keyboard selection, in keyboard list, when gdm starts in a livecd? Best regards, Pedro Silva -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: Esta ? uma parte de mensagem assinada digitalmente URL: From katzj at redhat.com Wed Oct 8 13:11:37 2008 From: katzj at redhat.com (Jeremy Katz) Date: Wed, 08 Oct 2008 09:11:37 -0400 Subject: [Fedora-livecd-list] [PATCH] Fix a bug with getting groups with options (@foo --optional). In-Reply-To: <1223422964-1096-1-git-send-email-jkeating@redhat.com> References: <1223404745.754.36.camel@aglarond.local> <1223422964-1096-1-git-send-email-jkeating@redhat.com> Message-ID: <1223471497.19673.2.camel@aglarond.local> On Tue, 2008-10-07 at 16:42 -0700, Jesse Keating wrote: > New yum api (in 3.2.x) allows us to pass into selectGroups what set we > want from the group. The default is mandatory and default. > Use that unless we have options from kickstart for either > optional packages (IE everything in the group) or just the > default packages of the group. Applied, thanks Jeremy From katzj at fedoraproject.org Wed Oct 8 13:12:52 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Wed, 8 Oct 2008 13:12:52 +0000 (UTC) Subject: [Fedora-livecd-list] imgcreate/yuminst.py Message-ID: <20081008131252.58C02C046B@lists.fedorahosted.org> imgcreate/yuminst.py | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) New commits: commit 74a5bc4cb1bb392ce1f8b59e76ffade8f505ae9c Author: Jesse Keating Date: Tue Oct 7 16:42:44 2008 -0700 Fix a bug with getting groups with options (@foo --optional). New yum api (in 3.2.x) allows us to pass into selectGroups what set we want from the group. The default is mandatory and default. Use that unless we have options from kickstart for either optional packages (IE everything in the group) or just the default packages of the group. This also allows us to just get the right packages the first time rather than getting packages and then pruning/adding as necessary. Less trips through the yum stack, which is a good thing. diff --git a/imgcreate/yuminst.py b/imgcreate/yuminst.py index a7b04b6..b89a6ed 100644 --- a/imgcreate/yuminst.py +++ b/imgcreate/yuminst.py @@ -110,11 +110,14 @@ class LiveCDYum(yum.YumBase): logging.warn("No such package %s to remove" %(pkg,)) def selectGroup(self, grp, include = pykickstart.parser.GROUP_DEFAULT): - yum.YumBase.selectGroup(self, grp) + # default to getting mandatory and default packages from a group + # unless we have specific options from kickstart + package_types = ['mandatory', 'default'] if include == pykickstart.parser.GROUP_REQUIRED: - map(lambda p: self.deselectPackage(p), grp.default_packages.keys()) + package_types.remove('default') elif include == pykickstart.parser.GROUP_ALL: - map(lambda p: self.selectPackage(p), grp.optional_packages.keys()) + package_types.append('optional') + yum.YumBase.selectGroup(self, grp, group_package_types=package_types) def addRepository(self, name, url = None, mirrorlist = None): def _varSubstitute(option): From katzj at fedoraproject.org Wed Oct 8 13:19:10 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Wed, 8 Oct 2008 13:19:10 +0000 (UTC) Subject: [Fedora-livecd-list] 2 commits - imgcreate/kickstart.py livecd-tools.spec Message-ID: <20081008131910.6B705C046B@lists.fedorahosted.org> imgcreate/kickstart.py | 3 ++- livecd-tools.spec | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) New commits: commit b8087f6e52deb6ff3d2c3d445fc82cdbf9e85c39 Author: Jeremy Katz Date: Wed Oct 8 09:17:11 2008 -0400 Don't lazily import rhpl.keyboard Lazily importing rhpl.keyboard can have some odd chroot impacts, so stop doing so and also be sure that we explicitly require rhpl diff --git a/imgcreate/kickstart.py b/imgcreate/kickstart.py index 278ac72..d911925 100644 --- a/imgcreate/kickstart.py +++ b/imgcreate/kickstart.py @@ -23,6 +23,8 @@ import subprocess import time import logging +import rhpl.keyboard + import pykickstart.commands as kscommands import pykickstart.constants as ksconstants import pykickstart.errors as kserrors @@ -133,7 +135,6 @@ class KeyboardConfig(KickstartConfig): # should this impact the X keyboard config too? # or do we want to make X be able to do this mapping? # - import rhpl.keyboard k = rhpl.keyboard.Keyboard() if kskeyboard.keyboard: k.set(kskeyboard.keyboard) diff --git a/livecd-tools.spec b/livecd-tools.spec index c08e2dd..2df16df 100644 --- a/livecd-tools.spec +++ b/livecd-tools.spec @@ -20,6 +20,7 @@ Requires: squashfs-tools Requires: pykickstart >= 0.96 Requires: dosfstools >= 2.11-8 Requires: isomd5sum +Requires: rhpl %ifarch %{ix86} x86_64 Requires: syslinux %endif commit 86577e6c20e6ff4347530682955b7c898e6b9545 Author: Jeremy Katz Date: Wed Oct 8 09:16:54 2008 -0400 Bump yum version for new API usage diff --git a/livecd-tools.spec b/livecd-tools.spec index ab641b8..c08e2dd 100644 --- a/livecd-tools.spec +++ b/livecd-tools.spec @@ -14,7 +14,7 @@ BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-root Requires: util-linux Requires: coreutils Requires: e2fsprogs -Requires: yum >= 3.1.7 +Requires: yum >= 3.2.18 Requires: mkisofs Requires: squashfs-tools Requires: pykickstart >= 0.96 From orion at cora.nwra.com Wed Oct 8 22:02:56 2008 From: orion at cora.nwra.com (Orion Poplawski) Date: Wed, 08 Oct 2008 16:02:56 -0600 Subject: [Fedora-livecd-list] umount: device is busy / cannot unlink ks-script Message-ID: <48ED2E10.9070602@cora.nwra.com> I'm trying to build a livecd with a custom ks file and getting: /sbin/restorecon reset /root/.gconf/desktop/gnome/url-handlers/%gconf.xml context unconfined_u:object_r:admin_home_t:s0->system_u:object_r:admin_home_t:s0 umount: /export/data1/tmp/imgcreate-OUBckN/install_root: device is busy. (In some cases useful info about processes that use the device is found by lsof(8) or fuser(1)) ioctl: LOOP_CLR_FD: Device or resource busy Traceback (most recent call last): File "/usr/bin/livecd-creator", line 136, in sys.exit(main()) File "/usr/bin/livecd-creator", line 120, in main creator.configure() File "//usr/lib/python2.5/site-packages/imgcreate/creator.py", line 727, in configure self.__run_post_scripts() File "//usr/lib/python2.5/site-packages/imgcreate/creator.py", line 699, in __run_post_scripts os.unlink(path) OSError: [Errno 2] No such file or directory: '/export/data1/tmp/imgcreate-OUBckN/install_root/tmp/ks-script-Gs6hh4' [root at test data1]# fuser -m /export/data1/tmp/imgcreate-OUBckN/install_root [root at test data1]# lsof | grep install_root [root at test data1]# umount /export/data1/tmp/imgcreate-OUBckN/install_root umount: /export/data1/tmp/imgcreate-OUBckN/install_root: device is busy. (In some cases useful info about processes that use the device is found by lsof(8) or fuser(1)) ? -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion at cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com From orion at cora.nwra.com Thu Oct 9 02:11:47 2008 From: orion at cora.nwra.com (orion at cora.nwra.com) Date: Wed, 8 Oct 2008 20:11:47 -0600 (MDT) Subject: [Fedora-livecd-list] umount: device is busy / cannot unlink ks-script In-Reply-To: <48ED2E10.9070602@cora.nwra.com> References: <48ED2E10.9070602@cora.nwra.com> Message-ID: <2412.71.208.51.222.1223518307.squirrel@www.cora.nwra.com> > I'm trying to build a livecd with a custom ks file and getting: > > /sbin/restorecon reset > /root/.gconf/desktop/gnome/url-handlers/%gconf.xml context > unconfined_u:object_r:admin_home_t:s0->system_u:object_r:admin_home_t:s0 > > umount: /export/data1/tmp/imgcreate-OUBckN/install_root: device is busy. > > (In some cases useful info about processes that use > > the device is found by lsof(8) or fuser(1)) > > ioctl: LOOP_CLR_FD: Device or resource busy Retried with no %pre/%post and it works, so I guess I need to poke around there... - Orion From katzj at redhat.com Thu Oct 9 02:49:13 2008 From: katzj at redhat.com (Jeremy Katz) Date: Wed, 08 Oct 2008 22:49:13 -0400 Subject: [Fedora-livecd-list] umount: device is busy / cannot unlink ks-script In-Reply-To: <2412.71.208.51.222.1223518307.squirrel@www.cora.nwra.com> References: <48ED2E10.9070602@cora.nwra.com> <2412.71.208.51.222.1223518307.squirrel@www.cora.nwra.com> Message-ID: <1223520553.19673.152.camel@aglarond.local> On Wed, 2008-10-08 at 20:11 -0600, orion at cora.nwra.com wrote: > > I'm trying to build a livecd with a custom ks file and getting: [snip] > > ioctl: LOOP_CLR_FD: Device or resource busy > > Retried with no %pre/%post and it works, so I guess I need to poke around > there... Yeah, usually this is caused by running something in your %post that leaves files open. The most common things are starting a service and not stopping it or starting some other long-running thing and not having it finish before the %post exists. Jeremy From cannewilson at googlemail.com Thu Oct 9 08:34:43 2008 From: cannewilson at googlemail.com (Anne Wilson) Date: Thu, 9 Oct 2008 09:34:43 +0100 Subject: [Fedora-livecd-list] Fedora on eeep, from live-cd Message-ID: <200810090934.44371.cannewilson@googlemail.com> Some time back I installed F9 from a kde live-cd onto a pen-drive, for use with my eeepc. Once in, I created myself as a user, then removed the fedora user. However, on boot, it still looks for fedora. It doesn't interfere with my login, but is a nuisance, and not something I want strangers to see. Is there any way to stop this happening? Anne -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part. URL: From orion at cora.nwra.com Thu Oct 9 19:10:43 2008 From: orion at cora.nwra.com (Orion Poplawski) Date: Thu, 09 Oct 2008 13:10:43 -0600 Subject: [Fedora-livecd-list] umount: device is busy / cannot unlink ks-script In-Reply-To: <1223520553.19673.152.camel@aglarond.local> References: <48ED2E10.9070602@cora.nwra.com> <2412.71.208.51.222.1223518307.squirrel@www.cora.nwra.com> <1223520553.19673.152.camel@aglarond.local> Message-ID: <48EE5733.2090300@cora.nwra.com> Jeremy Katz wrote: > On Wed, 2008-10-08 at 20:11 -0600, orion at cora.nwra.com wrote: >>> I'm trying to build a livecd with a custom ks file and getting: > [snip] >>> ioctl: LOOP_CLR_FD: Device or resource busy >> Retried with no %pre/%post and it works, so I guess I need to poke around >> there... > > Yeah, usually this is caused by running something in your %post that > leaves files open. The most common things are starting a service and > not stopping it or starting some other long-running thing and not having > it finish before the %post exists. Okay, this appears to be caused by: mount -t tmpfs varcacheyum /var/cache/yum mount -t tmpfs tmp /tmp mount -t tmpfs vartmp /var/tmp Which I took from /usr/share/fedora-release/fedora-live-base.ks. Is there a newer version of this file somewhere? Looking at fedora-release.spec there's not much of a clue if the sources are maintained anywhere else. URL: http://fedoraproject.org Source: %{name}-%{version}.tar.gz -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion at cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com From opensource at till.name Thu Oct 9 20:32:33 2008 From: opensource at till.name (Till Maas) Date: Thu, 09 Oct 2008 22:32:33 +0200 Subject: [Fedora-livecd-list] umount: device is busy / cannot unlink ks-script In-Reply-To: <48EE5733.2090300@cora.nwra.com> References: <48ED2E10.9070602@cora.nwra.com> <1223520553.19673.152.camel@aglarond.local> <48EE5733.2090300@cora.nwra.com> Message-ID: <200810092232.33989.opensource@till.name> On Thu October 9 2008, Orion Poplawski wrote: > Okay, this appears to be caused by: > > mount -t tmpfs varcacheyum /var/cache/yum Only a little bit related: You should probably add -o mode=0755 to the mount command here. Regards, Till -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 827 bytes Desc: This is a digitally signed message part. URL: From katzj at redhat.com Fri Oct 10 01:03:14 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 09 Oct 2008 21:03:14 -0400 Subject: [Fedora-livecd-list] umount: device is busy / cannot unlink ks-script In-Reply-To: <200810092232.33989.opensource@till.name> References: <48ED2E10.9070602@cora.nwra.com> <1223520553.19673.152.camel@aglarond.local> <48EE5733.2090300@cora.nwra.com> <200810092232.33989.opensource@till.name> Message-ID: <1223600594.19673.176.camel@aglarond.local> On Thu, 2008-10-09 at 22:32 +0200, Till Maas wrote: > On Thu October 9 2008, Orion Poplawski wrote: > > Okay, this appears to be caused by: > > > > mount -t tmpfs varcacheyum /var/cache/yum > > Only a little bit related: You should probably add -o mode=0755 to the mount > command here. Indeed, thanks -- fixed Jeremy From katzj at redhat.com Fri Oct 10 01:03:49 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 09 Oct 2008 21:03:49 -0400 Subject: [Fedora-livecd-list] umount: device is busy / cannot unlink ks-script In-Reply-To: <48EE5733.2090300@cora.nwra.com> References: <48ED2E10.9070602@cora.nwra.com> <2412.71.208.51.222.1223518307.squirrel@www.cora.nwra.com> <1223520553.19673.152.camel@aglarond.local> <48EE5733.2090300@cora.nwra.com> Message-ID: <1223600629.19673.177.camel@aglarond.local> On Thu, 2008-10-09 at 13:10 -0600, Orion Poplawski wrote: > Jeremy Katz wrote: > > On Wed, 2008-10-08 at 20:11 -0600, orion at cora.nwra.com wrote: > >>> I'm trying to build a livecd with a custom ks file and getting: > > [snip] > >>> ioctl: LOOP_CLR_FD: Device or resource busy > >> Retried with no %pre/%post and it works, so I guess I need to poke around > >> there... > > > > Yeah, usually this is caused by running something in your %post that > > leaves files open. The most common things are starting a service and > > not stopping it or starting some other long-running thing and not having > > it finish before the %post exists. > > Okay, this appears to be caused by: > > mount -t tmpfs varcacheyum /var/cache/yum > mount -t tmpfs tmp /tmp > mount -t tmpfs vartmp /var/tmp > > Which I took from /usr/share/fedora-release/fedora-live-base.ks. Is > there a newer version of this file somewhere? Looking at > fedora-release.spec there's not much of a clue if the sources are > maintained anywhere else. They used to be in the livecd-tools package/repo itself and just copied into fedora-release for some convenience. Now they are in the spin-kickstarts repo on fedorahosted Jeremy From katzj at fedoraproject.org Fri Oct 10 19:54:23 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Fri, 10 Oct 2008 19:54:23 +0000 (UTC) Subject: [Fedora-livecd-list] Changes to 'refs/tags/livecd-tools-019' Message-ID: <20081010195423.4BCD6C07B8@lists.fedorahosted.org> Tag 'livecd-tools-019' created by Jeremy Katz at 2008-10-10 14:23 +0000 livecd-tools-019 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEABECAAYFAkjvZX4ACgkQbZmr36XllQ2ACACgx/HdLcNqsti/oqqRZxcRMyMZ 0i0An0Buu2GqS+3SbPRiFmq9Wj2ZeFgM =xXvx -----END PGP SIGNATURE----- Changes since livecd-tools-018-17: --- 0 files changed --- From katzj at fedoraproject.org Fri Oct 10 19:54:21 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Fri, 10 Oct 2008 19:54:21 +0000 (UTC) Subject: [Fedora-livecd-list] 3 commits - imgcreate/yuminst.py livecd-tools.spec Makefile tools/livecd-iso-to-disk.sh Message-ID: <20081010195421.35312C046B@lists.fedorahosted.org> Makefile | 2 +- imgcreate/yuminst.py | 10 ++++++++-- livecd-tools.spec | 14 +++++++++++++- tools/livecd-iso-to-disk.sh | 2 +- 4 files changed, 23 insertions(+), 5 deletions(-) New commits: commit 0354c22595f761c8046767be816a281c6e288ae8 Author: Jeremy Katz Date: Fri Oct 10 10:23:31 2008 -0400 Bump version. diff --git a/Makefile b/Makefile index 05993c6..a463fc5 100644 --- a/Makefile +++ b/Makefile @@ -1,5 +1,5 @@ -VERSION = 018 +VERSION = 019 INSTALL = /usr/bin/install -c INSTALL_PROGRAM = ${INSTALL} diff --git a/livecd-tools.spec b/livecd-tools.spec index 2df16df..681f6cf 100644 --- a/livecd-tools.spec +++ b/livecd-tools.spec @@ -4,7 +4,7 @@ Summary: Tools for building live CD's Name: livecd-tools -Version: 018 +Version: 019 Release: 1%{?dist} License: GPLv2 Group: System Environment/Base @@ -64,6 +64,18 @@ rm -rf $RPM_BUILD_ROOT %{python_sitelib}/imgcreate/*.pyc %changelog +* Fri Oct 10 2008 Jeremy Katz - 019-1 +- livecd-iso-to-disk: Various other XO fixes +- Cleanup rpmdb locks after package installation +- Fix traceback due to lazy rhpl.keyboard import +- Fix using groups with options (jkeating) +- Support persistent /home on XO's internal flash +- Fix ramdisk load addr in boot/olpc.fth for XO +- Fix up boot from SD +- Fix extracting boot parameters for pxe (apevec) +- Make rpm macro information persist into the image (bkearney) +- Support %packages --instLangs (bkearney) + * Thu Aug 28 2008 Jeremy Katz - 018-1 - Use logging API for debugging messages (dhuff) - Some initial support for booting live images on an XO commit c80a302bda3cc1fc55fac088613bbaa86467fe86 Author: Jeremy Katz Date: Fri Oct 10 10:18:19 2008 -0400 Clean up rpmdb locks after installing packages also With new versions of RPM, people can otherwise run into things they don't expect inside of their %post scripts. So make things simpler for them diff --git a/imgcreate/yuminst.py b/imgcreate/yuminst.py index b89a6ed..04c9b40 100644 --- a/imgcreate/yuminst.py +++ b/imgcreate/yuminst.py @@ -69,10 +69,15 @@ class LiveCDYum(yum.YumBase): os.chmod(confpath, 0644) - def setup(self, confpath, installroot): - self._writeConf(confpath, installroot) + def _cleanupRpmdbLocks(self, installroot): + # cleans up temporary files left by bdb so that differing + # versions of rpm don't cause problems for f in glob.glob(installroot + "/var/lib/rpm/__db*"): os.unlink(f) + + def setup(self, confpath, installroot): + self._writeConf(confpath, installroot) + self._cleanupRpmdbLocks(installroot) self.doConfigSetup(fn = confpath, root = installroot) self.conf.cache = 0 self.doTsSetup() @@ -186,4 +191,5 @@ class LiveCDYum(yum.YumBase): cb.filelog = False ret = self.runTransaction(cb) print "" + self._cleanupRpmdbLocks(self.conf.installroot) return ret commit 742a7c5a1806362a927c70ed4cd79ddfb29ce9c4 Author: Jeremy Katz Date: Thu Oct 9 00:11:56 2008 -0400 Fix typo with --extra-kernel-args diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index f385dcb..d48f922 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -444,7 +444,7 @@ fi echo "Updating boot config file" # adjust label and fstype sed -i -e "s/CDLABEL=[^ ]*/$USBLABEL/" -e "s/rootfstype=[^ ]*/rootfstype=$USBFS/" $BOOTCONFIG -if [ -n "$kernelargs" ]; then sed -i -e "s/liveimg/liveimg ${kernelargs}/" $BOOTCONFIG.cfg ; fi +if [ -n "$kernelargs" ]; then sed -i -e "s/liveimg/liveimg ${kernelargs}/" $BOOTCONFIG ; fi if [ -n "$overlaysizemb" ]; then echo "Initializing persistent overlay file" From sundaram at fedoraproject.org Fri Oct 10 19:56:21 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Sat, 11 Oct 2008 01:26:21 +0530 Subject: [Fedora-livecd-list] 3 commits - imgcreate/yuminst.py livecd-tools.spec Makefile tools/livecd-iso-to-disk.sh In-Reply-To: <20081010195421.35312C046B@lists.fedorahosted.org> References: <20081010195421.35312C046B@lists.fedorahosted.org> Message-ID: <48EFB365.6010902@fedoraproject.org> Jeremy Katz wrote: > > commit c80a302bda3cc1fc55fac088613bbaa86467fe86 > Author: Jeremy Katz > Date: Fri Oct 10 10:18:19 2008 -0400 > > Clean up rpmdb locks after installing packages also > > With new versions of RPM, people can otherwise run into things they don't > expect inside of their %post scripts. So make things simpler for them Can you explain what you mean "things they don't expect" ? Not sure what problem is being solved here. Rahul From katzj at redhat.com Fri Oct 10 20:00:50 2008 From: katzj at redhat.com (Jeremy Katz) Date: Fri, 10 Oct 2008 16:00:50 -0400 Subject: [Fedora-livecd-list] 3 commits - imgcreate/yuminst.py livecd-tools.spec Makefile tools/livecd-iso-to-disk.sh In-Reply-To: <48EFB365.6010902@fedoraproject.org> References: <20081010195421.35312C046B@lists.fedorahosted.org> <48EFB365.6010902@fedoraproject.org> Message-ID: <1223668850.12340.8.camel@aglarond.local> On Sat, 2008-10-11 at 01:26 +0530, Rahul Sundaram wrote: > Jeremy Katz wrote: > > commit c80a302bda3cc1fc55fac088613bbaa86467fe86 > > Author: Jeremy Katz > > Date: Fri Oct 10 10:18:19 2008 -0400 > > > > Clean up rpmdb locks after installing packages also > > > > With new versions of RPM, people can otherwise run into things they don't > > expect inside of their %post scripts. So make things simpler for them > > Can you explain what you mean "things they don't expect" ? Not sure what > problem is being solved here. If you built an image that had a different version of rpm inside the chroot vs out (eg, say building a Fedora 9 live image on a rawhide box or vice versa) you end up with rpm telling you that the db versions don't match. And since we can safely know that the rpmdb isn't being accessed at that point, it is safe to remove /var/lib/rpm/__db* which are the lock files Jeremy From sundaram at fedoraproject.org Fri Oct 10 20:08:17 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Sat, 11 Oct 2008 01:38:17 +0530 Subject: [Fedora-livecd-list] 3 commits - imgcreate/yuminst.py livecd-tools.spec Makefile tools/livecd-iso-to-disk.sh In-Reply-To: <1223668850.12340.8.camel@aglarond.local> References: <20081010195421.35312C046B@lists.fedorahosted.org> <48EFB365.6010902@fedoraproject.org> <1223668850.12340.8.camel@aglarond.local> Message-ID: <48EFB631.6090108@fedoraproject.org> Jeremy Katz wrote: > If you built an image that had a different version of rpm inside the > chroot vs out (eg, say building a Fedora 9 live image on a rawhide box > or vice versa) Is that being supported by livecd-tools? I remember having all sort of issues including SELinux policy when trying to build a image where the host does not match what it inside the chroot (ie) building RHEL 5 images when the host in rawhide or anything like that. So now, I have VM's where the host roughly matches what is inside the chroot to avoid problems. Rahul From katzj at redhat.com Fri Oct 10 21:18:33 2008 From: katzj at redhat.com (Jeremy Katz) Date: Fri, 10 Oct 2008 17:18:33 -0400 Subject: [Fedora-livecd-list] 3 commits - imgcreate/yuminst.py livecd-tools.spec Makefile tools/livecd-iso-to-disk.sh In-Reply-To: <48EFB631.6090108@fedoraproject.org> References: <20081010195421.35312C046B@lists.fedorahosted.org> <48EFB365.6010902@fedoraproject.org> <1223668850.12340.8.camel@aglarond.local> <48EFB631.6090108@fedoraproject.org> Message-ID: <1223673513.12340.10.camel@aglarond.local> On Sat, 2008-10-11 at 01:38 +0530, Rahul Sundaram wrote: > Jeremy Katz wrote: > > If you built an image that had a different version of rpm inside the > > chroot vs out (eg, say building a Fedora 9 live image on a rawhide box > > or vice versa) > > Is that being supported by livecd-tools? I remember having all sort of > issues including SELinux policy when trying to build a image where the > host does not match what it inside the chroot (ie) building RHEL 5 > images when the host in rawhide or anything like that. So now, I have > VM's where the host roughly matches what is inside the chroot to avoid > problems. SELinux policy problems should be fixed these days as long as a) you're running 2.6.26+ as the host kernel so that you can support deferred unknown xattrs and b) there were some userspace fixes to "what's being installed" so that things wouldn't get confused. I think the only downside of not having this, though, was a longer relabel at the end Jeremy From dhuff at redhat.com Fri Oct 10 21:35:34 2008 From: dhuff at redhat.com (David Huff) Date: Fri, 10 Oct 2008 17:35:34 -0400 Subject: [Fedora-livecd-list] [PATCH] changed code to try to use os.rename first then shutil.move when moving around the images files. If the file are on the same fs will be alot faster. If not the rename fails and it reverts to shutil.move Message-ID: <1223674534-21642-1-git-send-email-dhuff@redhat.com> --- imgcreate/creator.py | 18 ++++++++++++++---- imgcreate/live.py | 16 ++++++++++++---- 2 files changed, 26 insertions(+), 8 deletions(-) diff --git a/imgcreate/creator.py b/imgcreate/creator.py index 519735e..72ab132 100644 --- a/imgcreate/creator.py +++ b/imgcreate/creator.py @@ -169,7 +169,10 @@ class ImageCreator(object): By default, this moves the install root into _outdir. """ - shutil.move(self._instroot, self._outdir + "/" + self.name) + try: + os.rename(self._instroot, self._outdir + "/" + self.name) + except Exception, e: + shutil.move(self._instroot, self._outdir + "/" + self.name) def _get_required_packages(self): """Return a list of required packages. @@ -750,8 +753,12 @@ class ImageCreator(object): self._stage_final_image() for f in os.listdir(self._outdir): - shutil.move(os.path.join(self._outdir, f), - os.path.join(destdir, f)) + try: + os.rename(os.path.join(self._outdir, f), + os.path.join(destdir, f)) + except Exception, e: + shutil.move(os.path.join(self._outdir, f), + os.path.join(destdir, f)) def create(self): """Install, configure and package an image. @@ -932,4 +939,7 @@ class LoopImageCreator(ImageCreator): def _stage_final_image(self): self._resparse() - shutil.move(self._image, self._outdir + "/" + self.name + ".img") + try: + os.rename(self._image, self._outdir + "/" + self.name + ".img") + except Exception, e: + shutil.move(self._image, self._outdir + "/" + self.name + ".img") diff --git a/imgcreate/live.py b/imgcreate/live.py index 17d3037..eab7be5 100644 --- a/imgcreate/live.py +++ b/imgcreate/live.py @@ -268,12 +268,20 @@ class LiveImageCreatorBase(LoopImageCreator): self._image, minimal_size) if self.skip_compression: - shutil.move(self._image, self.__isodir + "/LiveOS/ext3fs.img") + try: + os.rename(self._image, self.__isodir + "/LiveOS/ext3fs.img") + except Execption, e: + shutil.move(self._image, self.__isodir + "/LiveOS/ext3fs.img") else: makedirs(os.path.join(os.path.dirname(self._image), "LiveOS")) - shutil.move(self._image, - os.path.join(os.path.dirname(self._image), - "LiveOS", "ext3fs.img")) + try: + os.rename(self._image, + os.path.join(os.path.dirname(self._image), + "LiveOS", "ext3fs.img")) + except Exception, e: + shutil.move(self._image, + os.path.join(os.path.dirname(self._image), + "LiveOS", "ext3fs.img")) mksquashfs(os.path.dirname(self._image), self.__isodir + "/LiveOS/squashfs.img") -- 1.5.5.1 From opensource at till.name Tue Oct 14 15:41:46 2008 From: opensource at till.name (Till Maas) Date: Tue, 14 Oct 2008 17:41:46 +0200 Subject: [Fedora-livecd-list] Snap1 i686 Live "about Fedora" triggers "yelp not found" Message-ID: <200810141741.47289.opensource@till.name> Hiyas, I am not sure where to report this. I believed to have read that fedorahosted/releng hosts the source files for creating the live images, but I did not find them in the trac browser. So where is it and where is the apropriate bug tracker? The bug I noticed follows: I booted the Snap1 i686 Live iso via usb and selected "about Fedora" in the menu. Then I get an error message that yelp is not found. Regards, Till -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 827 bytes Desc: This is a digitally signed message part. URL: From katzj at redhat.com Tue Oct 14 16:10:34 2008 From: katzj at redhat.com (Jeremy Katz) Date: Tue, 14 Oct 2008 12:10:34 -0400 Subject: [Fedora-livecd-list] Snap1 i686 Live "about Fedora" triggers "yelp not found" In-Reply-To: <200810141741.47289.opensource@till.name> References: <200810141741.47289.opensource@till.name> Message-ID: <1224000634.12340.68.camel@aglarond.local> On Tue, 2008-10-14 at 17:41 +0200, Till Maas wrote: > I am not sure where to report this. I believed to have read that > fedorahosted/releng hosts the source files for creating the live images, but > I did not find them in the trac browser. So where is it and where is the > apropriate bug tracker? The bug I noticed follows: Bugs against live images should be reported in bugzilla in general. There are components for at least some of the live images and a generic "LiveCD" bucket as well. That reminds me I need to make sure there's a bucket for each of the F10 live images, though. > I booted the Snap1 i686 Live iso via usb and selected "about Fedora" in the > menu. Then I get an error message that yelp is not found. This is a bug in the about-fedora.desktop file (from fedora-release-notes iirc) not having TryExec. I already saw that this was filed. And also, we're hoping to get yelp back onto the live images I believe Jeremy From katzj at redhat.com Tue Oct 14 16:37:54 2008 From: katzj at redhat.com (Jeremy Katz) Date: Tue, 14 Oct 2008 12:37:54 -0400 Subject: [Fedora-livecd-list] [PATCH] changed code to try to use os.rename first then shutil.move when moving around the images files. If the file are on the same fs will be alot faster. If not the rename fails and it reverts to shutil.move In-Reply-To: <1223674534-21642-1-git-send-email-dhuff@redhat.com> References: <1223674534-21642-1-git-send-email-dhuff@redhat.com> Message-ID: <1224002274.12340.79.camel@aglarond.local> First, can you please follow good practice and use a short (<= 80, preferably 72) characters for the first line of your commits and then put the longer version? This makes things look a lot better in gitweb, git log, git shortlog, etc.... Also, shutil.move already does this so it really shouldn't be needed here as well def move(src, dst): """Recursively move a file or directory to another location. If the destination is on our current filesystem, then simply use rename. Otherwise, copy src to the dst and then remove src. A lot more could be done here... A look at a mv.c shows a lot of the issues this implementation glosses over. """ try: os.rename(src, dst) except OSError: if os.path.isdir(src): if destinsrc(src, dst): raise Error, "Cannot move a directory '%s' into itself '%s'." % (src, dst) copytree(src, dst, symlinks=True) rmtree(src) else: copy2(src,dst) os.unlink(src) Jeremy From opensource at till.name Tue Oct 14 16:51:31 2008 From: opensource at till.name (Till Maas) Date: Tue, 14 Oct 2008 18:51:31 +0200 Subject: [Fedora-livecd-list] Snap1 i686 Live "about Fedora" triggers "yelp not found" In-Reply-To: <1224000634.12340.68.camel@aglarond.local> References: <200810141741.47289.opensource@till.name> <1224000634.12340.68.camel@aglarond.local> Message-ID: <200810141851.38126.opensource@till.name> On Tue October 14 2008, Jeremy Katz wrote: > Bugs against live images should be reported in bugzilla in general. > There are components for at least some of the live images and a generic > "LiveCD" bucket as well. That reminds me I need to make sure there's a > bucket for each of the F10 live images, though. Ah thank you, there are Components for some different LiveCDs. Regards, Till -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 827 bytes Desc: This is a digitally signed message part. URL: From bjs at redhat.com Wed Oct 15 21:01:23 2008 From: bjs at redhat.com (Bryan J Smith) Date: Wed, 15 Oct 2008 17:01:23 -0400 Subject: [Fedora-livecd-list] Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... Message-ID: <1224104483.18461.29.camel@localhost.localdomain> First off, sorry for the cross-post. Secondly, let's take provisioning (Cobbler/Koan) and configuration management (Puppet) completely out of the question here. As much as I would love to focus on such, I'm looking at the boot/initrd aspects for now. I'd love to come back to them, but I'm on a tight schedule. Third, I've spent the last few days going through all versions of Anaconda, PyKickstart, Netboot and LiveCD (in addition to some LTSP prior, although I need to refresh myself) -- for RHEL 5 w/EPEL, the CentOS/Scientific "upgraded" versions (over RHEL 5/EPEL), Fedora 9 and even Fedora 10 Beta. I've really pulled out some hair. ;) Now more than ever it's clear to me that there have been many divergent paths taken, although I could just be ignorant (or stating that from ignorance). For this post, I will focus on netboot/disklessrc and livecd/mayflower, and possibly a replacement that focuses on base RHEL 5 support. Please correct any analysis, assumption, comment, etc... that is made in error or is otherwise leading down the wrong track. - Netboot/disklessrc The current approach in system-config-netboot for Diskless clients is legacy (not even nash, so it looks like some re-use from circa-2003 or even earlier), but the "disklessrc" approach with they way it loads in the Diskless initrd "just works." Not only does it "just work," but it works for RHEL 5 (possibly earlier). One of the complaints I've heard second-hand about supporting RHEL 5 in various setups (LTSP, etc...) is the lack of network card detection and setup, things that newer Anaconda and init processes seem to do in newer Fedora releases. Well, in all honesty, the disklessrc seems to do the job fine for RHEL 5 -- using PCI IDs, etc... Sure, it's not flawless, but for a diskless client (which is typically wired), it works. The "big thing" missing from the system-config-netboot is a lot of automation in generation of the system image/tree. E.g., it doesn't run Anaconda's install-to-disc like LiveCD does (including the EPEL 013-5 release for RHEL 5). - LiveCD/mayflower I've looked at the LiveCD tools now, including the mayflower initrd builder. I've also looked at the results. A lot of newer developments make assumptions on newer Anaconda/PyKickstart releases of post-RHEL 5 Fedora releases. I would definitely be interested in anyone who has concise information on what features are being used, otherwise, I'm looking at them as non-requirements (with my focus on RHEL 5). I've noted a few things I really don't care for, like dumping all modules in /lib/modules of the initrd, instead of maintaining some sort of tree like netboot/disklessrc does/uses. I also don't know if just pulling out the squashfs image is the most ideal for iso-to-disc, but I understand why it was done that way (to use the existing build). So it's clear it shouldn't be too terribly difficult to come up with something that adds in the Diskless option, network modules, etc... to add that. - What I want (should force myself to do) What I want -- aka what I should force myself to do so I more than a fat guy just eating up oxygen ;) -- is try to build a new, flexible initrd creator (call it "snitrd" out of my frustration ;) and . I don't know if there are other projects out there trying to do this, so please let me know if someone already has been tackling this. The concept is basically this (putting the Anaconda install-to-disc to the side for now) ... - Can generate a Diskless initrd (disklessrc init, various network) - Can generate a ISO initrd (like mayflower's init, various fs/etc...) - Can generate a Disk initrd (more than just squashfs? new init) - Can generate a "Super" initrd (take various boot init= options) Additionally, I was going to add the following options in the new "Disk" initrd init (think flash/embedded): - Option to compare image checksums to TFTP server, update if different - Option to boot diskless if local boot fails (trap failures/set flag) Does anyone else think this is a good idea? I don't think the Diskless, ISO and basic Disk initrd/init options are too difficult. The more embedded options could come later, as could the RHEL 5 Anaconda/PyKickstart v. newer Fedora differences/detections as well. Right now I could really use a single toolset to do at least some of the above. And I'm having to choose between the limited/stagnant Netboot/disklessrc and the more leading Fedora-centric LiveCD/mayflower. Anyone think I'm off my rocker? Over-complicating a simple issue? Don't know the first thing of what I'm talking about? Etc...? ;) -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From todd655495 at yahoo.com Wed Oct 15 22:32:34 2008 From: todd655495 at yahoo.com (Todd N) Date: Wed, 15 Oct 2008 15:32:34 -0700 (PDT) Subject: [Fedora-livecd-list] error reading information on service xxxx Message-ID: <191031.84222.qm@web54004.mail.re2.yahoo.com> Hello, ? During the boot process for a Live CD I have created with livecd-creator, I am getting error messages of the format ? error reading information on service xxxxxx: No such file or directory ? I am getting 6 such messages.? The services listed are bluetooth, isdn, rpcgssd, rpcidmapd, rpcsvcgssd, and sendmail.? None of these packages appear to be installed or running on my system.? I'd like to get rid of these messages if possible.? Any thoughts? ? Thanks much, Todd -------------- next part -------------- An HTML attachment was scrubbed... URL: From notting at redhat.com Thu Oct 16 14:11:52 2008 From: notting at redhat.com (Bill Nottingham) Date: Thu, 16 Oct 2008 10:11:52 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224104483.18461.29.camel@localhost.localdomain> References: <1224104483.18461.29.camel@localhost.localdomain> Message-ID: <20081016141152.GB5834@nostromo.devel.redhat.com> Bryan J Smith (bjs at redhat.com) said: > - Netboot/disklessrc > > The current approach in system-config-netboot for Diskless clients is > legacy (not even nash, so it looks like some re-use from circa-2003 or > even earlier), but the "disklessrc" approach with they way it loads in > the Diskless initrd "just works." Not only does it "just work," but it > works for RHEL 5 (possibly earlier). This code is dead (long-term.) Note that mkinitrd itself has support for NFS boot, iSCSI, etc. > - LiveCD/mayflower > > I've looked at the LiveCD tools now, including the mayflower initrd > builder. I've also looked at the results. A lot of newer developments > make assumptions on newer Anaconda/PyKickstart releases of post-RHEL 5 > Fedora releases. I would definitely be interested in anyone who has > concise information on what features are being used, otherwise, I'm > looking at them as non-requirements (with my focus on RHEL 5). There is a livecd-tools branch for RHEL 5 - I believe Rahul maintains that. > I've noted a few things I really don't care for, like dumping all > modules in /lib/modules of the initrd, instead of maintaining some sort > of tree like netboot/disklessrc does/uses. Attempting to catalog a list of network/storage module is a long-term failure. We do a hack in anaconda by looking for modules that reference certain needed symbols (register_net_device, register_block, etc.), but it's still a hack. > - What I want (should force myself to do) > > What I want -- aka what I should force myself to do so I more than a fat > guy just eating up oxygen ;) -- is try to build a new, flexible initrd > creator (call it "snitrd" out of my frustration ;) and . I don't know > if there are other projects out there trying to do this, so please let > me know if someone already has been tackling this. > > The concept is basically this (putting the Anaconda install-to-disc to > the side for now) ... > > - Can generate a Diskless initrd (disklessrc init, various network) > - Can generate a ISO initrd (like mayflower's init, various fs/etc...) > - Can generate a Disk initrd (more than just squashfs? new init) > - Can generate a "Super" initrd (take various boot init= options) > > Additionally, I was going to add the following options in the new "Disk" > initrd init (think flash/embedded): > > - Option to compare image checksums to TFTP server, update if different > - Option to boot diskless if local boot fails (trap failures/set flag) > > Does anyone else think this is a good idea? I don't think the Diskless, > ISO and basic Disk initrd/init options are too difficult. The more > embedded options could come later, as could the RHEL 5 > Anaconda/PyKickstart v. newer Fedora differences/detections as well. > > Right now I could really use a single toolset to do at least some of the > above. And I'm having to choose between the limited/stagnant > Netboot/disklessrc and the more leading Fedora-centric LiveCD/mayflower. > > Anyone think I'm off my rocker? Over-complicating a simple issue? > Don't know the first thing of what I'm talking about? Etc...? ;) Honestly, I'm not sure what we really need right now is Yet Another initrd project. There's talk of a new upstream one that can hopefully be shared by multiple distros, but that's still a ways off from production. In the meantime, it's probably better to just fix what's there. Bill From bjs at redhat.com Thu Oct 16 14:30:57 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 10:30:57 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <20081016141152.GB5834@nostromo.devel.redhat.com> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> Message-ID: <1224167457.4111.13.camel@localhost.localdomain> On Thu, 2008-10-16 at 10:11 -0400, Bill Nottingham wrote: > This code is dead (long-term.) Note that mkinitrd itself has support > for NFS boot, iSCSI, etc. Understand that, but there is still a modification of NFS boot for Diskless/Stateless in RHEL 5, even if based on older developments. > There is a livecd-tools branch for RHEL 5 - I believe Rahul maintains > that. Yes, that's what I'm looking at. It's livecd-tools 013-5 in EPEL. I've discussed with him off-list. > Attempting to catalog a list of network/storage module is a long-term > failure. We do a hack in anaconda by looking for modules that reference > certain needed symbols (register_net_device, register_block, etc.), but > it's still a hack. And only in newer Fedora developments (post-RHEL 5) as I understand it? > Honestly, I'm not sure what we really need right now is Yet Another > initrd project. There's talk of a new upstream one that can hopefully > be shared by multiple distros, but that's still a ways off from production. > In the meantime, it's probably better to just fix what's there. Which means merging system-config-netboot approaches, its init script, etc... with mayflower for RHEL 5-specific details that do not have newer Fedora developments. Furthermore, I guess the "initrd" was the inappropriate term to use. What I'm talking more about is radical departure from what "mayflower" currently does, to look at the problem more generically -- not just for LiveCDs, but Diskless, Flash, etc... -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From notting at redhat.com Thu Oct 16 14:41:49 2008 From: notting at redhat.com (Bill Nottingham) Date: Thu, 16 Oct 2008 10:41:49 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224167457.4111.13.camel@localhost.localdomain> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> Message-ID: <20081016144149.GE5834@nostromo.devel.redhat.com> Bryan J Smith (bjs at redhat.com) said: > Understand that, but there is still a modification of NFS boot for > Diskless/Stateless in RHEL 5, even if based on older developments. Not sure what you mean by 'a modification'. The support *is there in the stock initrd in RHEL 5*. > > Attempting to catalog a list of network/storage module is a long-term > > failure. We do a hack in anaconda by looking for modules that reference > > certain needed symbols (register_net_device, register_block, etc.), but > > it's still a hack. > > And only in newer Fedora developments (post-RHEL 5) as I understand it? Correct. It was a side effect of porting the installer to udev. > > Honestly, I'm not sure what we really need right now is Yet Another > > initrd project. There's talk of a new upstream one that can hopefully > > be shared by multiple distros, but that's still a ways off from production. > > In the meantime, it's probably better to just fix what's there. > > Which means merging system-config-netboot approaches, its init script, > etc... with mayflower for RHEL 5-specific details that do not have newer > Fedora developments. What specific features do you need? Bill From bjs at redhat.com Thu Oct 16 16:08:10 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 12:08:10 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <20081016144149.GE5834@nostromo.devel.redhat.com> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> Message-ID: <1224173290.4111.20.camel@localhost.localdomain> On Thu, 2008-10-16 at 10:41 -0400, Bill Nottingham wrote: > Not sure what you mean by 'a modification'. The support *is there > in the stock initrd in RHEL 5*. Support is there, yes. But for diskless to work, or should I say, as system-config-netboot modifies, it adds modules, a flexible disklessrc init and related support files. Again, it was my mistake to refer to this as an "initrd" modification. I guess what I'm talking about is a replacement "mayflower" that does many other things. > Correct. It was a side effect of porting the installer to udev. Which leaves me with older developments, Rahul's modification of an older LiveCD/mayflower, etc... > What specific features do you need? As my original post, a single toolset build around Anaconda install-to-disk, like LiveCD/mayflower, but for RHEL 5 today (will adapt to newer Fedora developments for a future RHEL 6 when I have more time) ... - Can generate a Diskless initrd (disklessrc init, various network) - Can generate a ISO initrd (like mayflower's init, various fs/etc...) - Can generate a Disk initrd (more than just squashfs? new init) - Can generate a "Super" initrd (take various boot init= options) Additionally, I was going to add the following options in the new "Disk" initrd init (think flash/embedded): - Option to compare image checksums to TFTP server, update if different - Option to boot diskless if local boot fails (trap failures/set flag) -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From notting at redhat.com Thu Oct 16 16:12:43 2008 From: notting at redhat.com (Bill Nottingham) Date: Thu, 16 Oct 2008 12:12:43 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224173290.4111.20.camel@localhost.localdomain> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> Message-ID: <20081016161243.GB23009@nostromo.devel.redhat.com> Bryan J Smith (bjs at redhat.com) said: > On Thu, 2008-10-16 at 10:41 -0400, Bill Nottingham wrote: > > Not sure what you mean by 'a modification'. The support *is there > > in the stock initrd in RHEL 5*. > > Support is there, yes. But for diskless to work, or should I say, as > system-config-netboot modifies, it adds modules, a flexible disklessrc > init and related support files. ... because the original developer made the mistake of not working with the current mkinitrd first. > > What specific features do you need? > > As my original post, a single toolset build around Anaconda > install-to-disk, like LiveCD/mayflower, but for RHEL 5 today (will adapt > to newer Fedora developments for a future RHEL 6 when I have more > time) ... > > - Can generate a Diskless initrd (disklessrc init, various network) > - Can generate a ISO initrd (like mayflower's init, various fs/etc...) > - Can generate a Disk initrd (more than just squashfs? new init) > - Can generate a "Super" initrd (take various boot init= options) Why not work with backporting existing initrd features, then? The mayflower stuff was merged in in current Fedora. Bill From bjs at redhat.com Thu Oct 16 16:19:32 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 12:19:32 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <20081016161243.GB23009@nostromo.devel.redhat.com> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> Message-ID: <1224173972.4111.26.camel@localhost.localdomain> On Thu, 2008-10-16 at 12:12 -0400, Bill Nottingham wrote: > ... because the original developer made the mistake of not working with > the current mkinitrd first. Then I'd like to very much correct that, for RHEL 5 to start. ;) I do note it's rather legacy. The init is even bash, not nash. If you have recommendations, now's the time to make them. ;) > Why not work with backporting existing initrd features, then? It seems there's quite a bit of assumptions around newer Anaconda releases that RHEL 5 doesn't ship. I want to say that's the same complaint in getting the LTSP client working on RHEL 5 -- the lack of various init shipped with newer Fedora/Anaconda releases that RHEL 5 lacks. If I was working on RHEL 6 stuff, I'm there! But I'm not (let alone my schedule did not budget for this). If you have notes, recommendations, etc... that would let me modify it in short order, I'm all ears. That and how to bring those features into mayflower in a way that works for RHEL 5 Anaconda, PyKickstart, init-scripts, etc... ;) > The mayflower stuff was merged in in current Fedora. I still see it in Rahul's livecd-tools 013-5 in EPEL for RHEL 5. -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From sundaram at fedoraproject.org Thu Oct 16 16:42:54 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Thu, 16 Oct 2008 22:12:54 +0530 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224173972.4111.26.camel@localhost.localdomain> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> Message-ID: <48F76F0E.4050602@fedoraproject.org> Bryan J Smith wrote: >> The mayflower stuff was merged in in current Fedora. > > I still see it in Rahul's livecd-tools 013-5 in EPEL for RHEL 5. For one, nobody in EPEL can control what gets merged into the core product which leaves room for us just enough to workaround the things we really need. Any movement beyond that is pain. What gets merged into Fedora however makes it future proof to the next RHEL version. Rahul From jkeating at redhat.com Thu Oct 16 16:44:39 2008 From: jkeating at redhat.com (Jesse Keating) Date: Thu, 16 Oct 2008 09:44:39 -0700 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224173972.4111.26.camel@localhost.localdomain> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> Message-ID: <1224175479.4122.76.camel@luminos.localdomain> On Thu, 2008-10-16 at 12:19 -0400, Bryan J Smith wrote: > I do note it's rather legacy. The init is even bash, not nash. If you > have recommendations, now's the time to make them. ;) Here's one, that's not going to be very popular. Stop trying to pile features onto a platform that is supposed to be stable. -- Jesse Keating Fedora -- Freedom? is a feature! identi.ca: http://identi.ca/jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part URL: From bjs at redhat.com Thu Oct 16 16:48:41 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 12:48:41 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224175479.4122.76.camel@luminos.localdomain> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> Message-ID: <1224175721.4111.29.camel@localhost.localdomain> On Thu, 2008-10-16 at 09:44 -0700, Jesse Keating wrote: > Here's one, that's not going to be very popular. Stop trying to pile > features onto a platform that is supposed to be stable. So what you're saying is that I should just continue to use "system-config-netboot" and "livecd-tools" as-is? Not going to fly here for a thin client project. ;) -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From bjs at redhat.com Thu Oct 16 16:50:19 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 12:50:19 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <48F76F0E.4050602@fedoraproject.org> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <48F76F0E.4050602@fedoraproject.org> Message-ID: <1224175819.4111.32.camel@localhost.localdomain> On Thu, 2008-10-16 at 22:12 +0530, Rahul Sundaram wrote: > For one, nobody in EPEL can control what gets merged into the core > product which leaves room for us just enough to workaround the things we > really need. Which I'm also trying to avoid. I don't want to use any updated Anaconda, PyKickstart, etc... > Any movement beyond that is pain. What gets merged into > Fedora however makes it future proof to the next RHEL version. Understand. But my requirements are very much RHEL 5. I'll focus on RHEL 6 when I'm passed this. ;) So, your thoughts on my mayflower replacement concept? -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From jkeating at redhat.com Thu Oct 16 16:56:57 2008 From: jkeating at redhat.com (Jesse Keating) Date: Thu, 16 Oct 2008 09:56:57 -0700 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224175721.4111.29.camel@localhost.localdomain> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> Message-ID: <1224176217.4122.78.camel@luminos.localdomain> On Thu, 2008-10-16 at 12:48 -0400, Bryan J Smith wrote: > > So what you're saying is that I should just continue to use > "system-config-netboot" and "livecd-tools" as-is? > > Not going to fly here for a thin client project. ;) All I'm saying is that trying to pile stuff that's designed for post RHEL5 into RHEL5 is not going to very conducive to that "stability" of RHEL5. Apart from that, you're not likely to find much help in the various upstreams trying to backport those things to the old RHEL5 platform. And it is quite old. -- Jesse Keating Fedora -- Freedom? is a feature! identi.ca: http://identi.ca/jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part URL: From notting at redhat.com Thu Oct 16 16:59:57 2008 From: notting at redhat.com (Bill Nottingham) Date: Thu, 16 Oct 2008 12:59:57 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224175721.4111.29.camel@localhost.localdomain> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> Message-ID: <20081016165957.GA23397@nostromo.devel.redhat.com> Bryan J Smith (bjs at redhat.com) said: > On Thu, 2008-10-16 at 09:44 -0700, Jesse Keating wrote: > > Here's one, that's not going to be very popular. Stop trying to pile > > features onto a platform that is supposed to be stable. > > So what you're saying is that I should just continue to use > "system-config-netboot" and "livecd-tools" as-is? I'm speaking from a bit of a position of prejudice, but s-c-netboot is almost never the right tool. But in general, I think you're in one-off land - I don't think writing a whole new thing and expecting to forward port it is going to be worth the effort. Bill From bjs at redhat.com Thu Oct 16 17:07:55 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 13:07:55 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <20081016165957.GA23397@nostromo.devel.redhat.com> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> <20081016165957.GA23397@nostromo.devel.redhat.com> Message-ID: <1224176875.10074.5.camel@localhost.localdomain> On Thu, 2008-10-16 at 12:59 -0400, Bill Nottingham wrote: > I'm speaking from a bit of a position of prejudice, but > s-c-netboot is almost never the right tool. I would not disagree. But it did give me some "hints," at least what resulted. I was considering writing a partial replacement, just for the Diskless and Thin Client portions. On Thu, 2008-10-16 at 12:59 -0400, Bill Nottingham wrote: > But in general, I think you're in one-off land - I don't think > writing a whole new thing and expecting to forward port it i > going to be worth the effort. Maybe not. But I need something working for RHEL 5. If I can reuse some things (like the flash boot+checker), then great. If not, oh well. On Thu, 2008-10-16 at 09:56 -0700, Jesse Keating wrote: > All I'm saying is that trying to pile stuff that's designed for post > RHEL5 into RHEL5 is not going to very conducive to that "stability" of > RHEL5. Agreed. Which is why I just wanted to take "what works" in the "Stateless Tech Preview" in RHEL 5.2 (5.1+) and make a more LiveCD/mayflower-like approach. Use Anaconda to install to a directory, then post-process the initrd for Diskless, disk (including boot+update check), ISO, etc... On Thu, 2008-10-16 at 09:56 -0700, Jesse Keating wrote: > Apart from that, you're not likely to find much help in the > various upstreams trying to backport those things to the old RHEL5 > platform. And it is quite old. Exactomundo. I'm looking at, essentially, Fedora Core 6 (maybe a few Fedora 7 things if I'm lucky). Which brings me back to my "I need a solution, and I'll hack it out." So I'm just looking for comments, etc... I wish RHEL 6 was even in pre-Beta, but I do not believe it is. -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From notting at redhat.com Thu Oct 16 17:11:57 2008 From: notting at redhat.com (Bill Nottingham) Date: Thu, 16 Oct 2008 13:11:57 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224176875.10074.5.camel@localhost.localdomain> References: <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> <20081016165957.GA23397@nostromo.devel.redhat.com> <1224176875.10074.5.camel@localhost.localdomain> Message-ID: <20081016171157.GF23397@nostromo.devel.redhat.com> Bryan J Smith (bjs at redhat.com) said: > Agreed. Which is why I just wanted to take "what works" in the > "Stateless Tech Preview" in RHEL 5.2 (5.1+) and make a more > LiveCD/mayflower-like approach. Use Anaconda to install to a directory, > then post-process the initrd for Diskless, disk (including boot+update > check), ISO, etc... For stateless? Obviously the issue here is no easy support for any client-side state specifics, although you may be to graft something in. Not to mention the issues of having the whole FS be pseudo-read-write, which can lead to memory exhaustion if you're not careful. > > Apart from that, you're not likely to find much help in the > > various upstreams trying to backport those things to the old RHEL5 > > platform. And it is quite old. > > Exactomundo. I'm looking at, essentially, Fedora Core 6 (maybe a few > Fedora 7 things if I'm lucky). Right, and since the livecd stuff was finished *after* that, there's no wonder you're running into issues. Bill From bjs at redhat.com Thu Oct 16 17:18:35 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 13:18:35 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <20081016171157.GF23397@nostromo.devel.redhat.com> References: <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> <20081016165957.GA23397@nostromo.devel.redhat.com> <1224176875.10074.5.camel@localhost.localdomain> <20081016171157.GF23397@nostromo.devel.redhat.com> Message-ID: <1224177515.10074.11.camel@localhost.localdomain> On Thu, 2008-10-16 at 13:11 -0400, Bill Nottingham wrote: > For stateless? Obviously the issue here is no easy support for > any client-side state specifics, although you may be to graft > something in. Not to mention the issues of having the whole > FS be pseudo-read-write, which can lead to memory exhaustion > if you're not careful. Yes, I'm well aware. > Right, and since the livecd stuff was finished *after* that, there's > no wonder you're running into issues. Which brings me back to Rahul's EPEL livecd-tools 013-5 versus the "Stateless Tech Preview" in RHEL 5.2 (really anything 5.1+). I see it's looking more like a merger of the two. I'm now looking at the mkinitrd of Fedora 9 versus RHEL 5. Is it work generating a patch (which will have obvious conditionals) against newer mkinitrd developments (such as the Fedora 10 versoin) so it can work for RHEL 5? Or would everyone agree it's probably not a good idea to try to work something for RHEL 5? My idea was to replace mayflower with something for RHEL 5 now, and then work on adding those features to mkinitrd for possibly RHEL 6 (or some future development) when I have time. I know this is not the only client that is going to want some of these options. -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From katzj at redhat.com Thu Oct 16 17:05:41 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 16 Oct 2008 13:05:41 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <20081016161243.GB23009@nostromo.devel.redhat.com> References: <1224104483.18461.29.camel@localhost.localdomain> <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> Message-ID: <1224176741.13313.2.camel@aglarond.local> On Thu, 2008-10-16 at 12:12 -0400, Bill Nottingham wrote: > Bryan J Smith (bjs at redhat.com) said: > > As my original post, a single toolset build around Anaconda > > install-to-disk, like LiveCD/mayflower, but for RHEL 5 today (will adapt > > to newer Fedora developments for a future RHEL 6 when I have more > > time) ... > > > > - Can generate a Diskless initrd (disklessrc init, various network) > > - Can generate a ISO initrd (like mayflower's init, various fs/etc...) > > - Can generate a Disk initrd (more than just squashfs? new init) > > - Can generate a "Super" initrd (take various boot init= options) > > Why not work with backporting existing initrd features, then? The > mayflower stuff was merged in in current Fedora. It's a bit of a stretch to really say that it was merged... there are more things in common but there's still a lot more divergence than I'd really like :/ Jeremy From katzj at redhat.com Thu Oct 16 17:32:37 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 16 Oct 2008 13:32:37 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224177515.10074.11.camel@localhost.localdomain> References: <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> <20081016165957.GA23397@nostromo.devel.redhat.com> <1224176875.10074.5.camel@localhost.localdomain> <20081016171157.GF23397@nostromo.devel.redhat.com> <1224177515.10074.11.camel@localhost.localdomain> Message-ID: <1224178357.370.3.camel@aglarond.local> On Thu, 2008-10-16 at 13:18 -0400, Bryan J Smith wrote: > On Thu, 2008-10-16 at 13:11 -0400, Bill Nottingham wrote: > > Right, and since the livecd stuff was finished *after* that, there's > > no wonder you're running into issues. > > Which brings me back to Rahul's EPEL livecd-tools 013-5 versus the > "Stateless Tech Preview" in RHEL 5.2 (really anything 5.1+). I see it's > looking more like a merger of the two. > > I'm now looking at the mkinitrd of Fedora 9 versus RHEL 5. Is it work > generating a patch (which will have obvious conditionals) against newer > mkinitrd developments (such as the Fedora 10 versoin) so it can work for > RHEL 5? I think I can pretty safely speak for Peter and say that it's not something that's interesting. We need to move forward, not keep navel gazing back at code from 2 years ago. Jeremy From katzj at redhat.com Thu Oct 16 17:39:59 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 16 Oct 2008 13:39:59 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224177515.10074.11.camel@localhost.localdomain> References: <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> <20081016165957.GA23397@nostromo.devel.redhat.com> <1224176875.10074.5.camel@localhost.localdomain> <20081016171157.GF23397@nostromo.devel.redhat.com> <1224177515.10074.11.camel@localhost.localdomain> Message-ID: <1224178799.370.9.camel@aglarond.local> Hrm. evo ate part of my reply so trying to recreate it.. On Thu, 2008-10-16 at 13:18 -0400, Bryan J Smith wrote: > Or would everyone agree it's probably not a good idea to try to work > something for RHEL 5? My idea was to replace mayflower with something > for RHEL 5 now, and then work on adding those features to mkinitrd for > possibly RHEL 6 (or some future development) when I have time. It's very much not the right idea to start doing the work on RHEL5. If there's functionality that people want, you want to start working on it against the upstream development tree. Not against a two year old branch. Then, once you get something which is working acceptably upstream and meets the requirements, it's reasonable to assess the feasibility of a backport. Doing it ANY OTHER WAY all but assures the fact that you'll end up having to do the work all over again against a new version. And thus the cycle ends up repeating itself endlessly. For an analogy -- if you were to start working on an entirely new kernel subsystem, would you do it against 2.6.18 or would you do it against Linus's git? Jeremy From bjs at redhat.com Thu Oct 16 18:26:48 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 14:26:48 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224178799.370.9.camel@aglarond.local> References: <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> <20081016165957.GA23397@nostromo.devel.redhat.com> <1224176875.10074.5.camel@localhost.localdomain> <20081016171157.GF23397@nostromo.devel.redhat.com> <1224177515.10074.11.camel@localhost.localdomain> <1224178799.370.9.camel@aglarond.local> Message-ID: <1224181608.10074.17.camel@localhost.localdomain> On Thu, 2008-10-16 at 13:18 -0400, Bryan J Smith wrote: > Or would everyone agree it's probably not a good idea to try to work > something for RHEL 5? First off, er um, this came out wrong. ;) What I mean to say is "would everyone agree anything I do for RHEL 5 is probably not going to be useful for anything else?" I'm stuck with RHEL 5 for my client, end of story there, until RHEL 6. ;) On Thu, 2008-10-16 at 13:39 -0400, Jeremy Katz wrote: > It's very much not the right idea to start doing the work on RHEL5. Sorry, re-read the previous (now corrected). My apologies for my prior wording. On Thu, 2008-10-16 at 13:39 -0400, Jeremy Katz wrote: > If there's functionality that people want, you want to start working on it > against the upstream development tree. Not against a two year old > branch. Then, once you get something which is working acceptably > upstream and meets the requirements, it's reasonable to assess the > feasibility of a backport. I wish I could start that way. On Thu, 2008-10-16 at 13:39 -0400, Jeremy Katz wrote: > Doing it ANY OTHER WAY all but assures the fact that you'll end up > having to do the work all over again against a new version. And thus > the cycle ends up repeating itself endlessly. > For an analogy -- if you were to start working on an entirely new > kernel subsystem, would you do it against 2.6.18 or would you do it > against Linus's git? I agree with you completely on this. Unfortunately, I don't have the time and focus. When I'm past this, I will certainly rebase for upstream -- especially for (and likely very prior) to RHEL 6. -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From bjs at redhat.com Thu Oct 16 18:27:55 2008 From: bjs at redhat.com (Bryan J Smith) Date: Thu, 16 Oct 2008 14:27:55 -0400 Subject: [Fedora-livecd-list] Re: Seemingly divergent netboot/disklessrc and livecd/mayflower for initrd, init, etc... In-Reply-To: <1224178357.370.3.camel@aglarond.local> References: <20081016141152.GB5834@nostromo.devel.redhat.com> <1224167457.4111.13.camel@localhost.localdomain> <20081016144149.GE5834@nostromo.devel.redhat.com> <1224173290.4111.20.camel@localhost.localdomain> <20081016161243.GB23009@nostromo.devel.redhat.com> <1224173972.4111.26.camel@localhost.localdomain> <1224175479.4122.76.camel@luminos.localdomain> <1224175721.4111.29.camel@localhost.localdomain> <20081016165957.GA23397@nostromo.devel.redhat.com> <1224176875.10074.5.camel@localhost.localdomain> <20081016171157.GF23397@nostromo.devel.redhat.com> <1224177515.10074.11.camel@localhost.localdomain> <1224178357.370.3.camel@aglarond.local> Message-ID: <1224181675.10074.19.camel@localhost.localdomain> On Thu, 2008-10-16 at 13:32 -0400, Jeremy Katz wrote: > I think I can pretty safely speak for Peter and say that it's not > something that's interesting. We need to move forward, not keep navel > gazing back at code from 2 years ago. That's what I figured. On Thu, 2008-10-16 at 13:05 -0400, Jeremy Katz wrote: > It's a bit of a stretch to really say that it was merged... there are > more things in common but there's still a lot more divergence than I'd > really like :/ Which is where I came from, after studying code for a few days. ;) -- Bryan J Smith - Senior Consultant - Red Hat GPS SE US mailto:bjs at redhat.com +1 (407) 489-7013 (Mobile) mailto:b.j.smith at ieee.org (non-RH/ext to Blackberry) ----------------------------------------------------- For every dollar you spend on Red Hat solutions, you not only fund the leading community development re- source, but you receive the #1 IT industry leader in corporate value. http://www.redhat.com/promo/vendor/ From bkearney at redhat.com Thu Oct 16 21:11:43 2008 From: bkearney at redhat.com (Bryan Kearney) Date: Thu, 16 Oct 2008 17:11:43 -0400 Subject: [Fedora-livecd-list] [PATCH] Allow for http:// or ftp:// urls to be passed in to the image creators Message-ID: <1224191503-7562-1-git-send-email-bkearney@redhat.com> From: root --- imgcreate/__init__.py | 2 ++ imgcreate/kickstart.py | 20 +++++++++++++++++++- tools/livecd-creator | 4 ++-- 3 files changed, 23 insertions(+), 3 deletions(-) diff --git a/imgcreate/__init__.py b/imgcreate/__init__.py index bcfb40e..7cf30b8 100644 --- a/imgcreate/__init__.py +++ b/imgcreate/__init__.py @@ -34,6 +34,7 @@ Also exported are: - CreatorError - all exceptions throw are of this type - FSLABEL_MAXLEN - the length to which LoopImageCreator.fslabel is truncated - read_kickstart() - a utility function for kickstart parsing + - verify_kickstart() - a utility function to verify the existence of a kickstart file - build_name() - a utility to construct an image name Each of the creator classes are designed to be subclassable, allowing the user @@ -63,6 +64,7 @@ __all__ = ( 'LoopImageCreator', 'FSLABEL_MAXLEN', 'read_kickstart', + 'verify_kickstart', 'construct_name', 'setup_logging', ) diff --git a/imgcreate/kickstart.py b/imgcreate/kickstart.py index d911925..9de347a 100644 --- a/imgcreate/kickstart.py +++ b/imgcreate/kickstart.py @@ -22,6 +22,7 @@ import shutil import subprocess import time import logging +import urllib import rhpl.keyboard @@ -34,6 +35,19 @@ import pykickstart.version as ksversion import imgcreate.errors as errors import imgcreate.fs as fs +def path_is_url(path): + pupper = path.upper() + return (pupper.startswith("HTTP://")) or (pupper.startswith("FTP://")) + +def verify_kickstart(path): + verified = False + if not path is None: + if path_is_url(path): + verified = True + else: + verified = os.path.isfile(path) + return verified + def read_kickstart(path): """Parse a kickstart file and return a KickstartParser instance. @@ -47,7 +61,11 @@ def read_kickstart(path): version = ksversion.makeVersion() ks = ksparser.KickstartParser(version) try: - ks.readKickstart(path) + if path_is_url(path): + ksfile,headers = urllib.urlretrieve(path) + ks.readKickstart(ksfile) + else: + ks.readKickstart(path) except IOError, (err, msg): raise errors.KickstartError("Failed to read kickstart file " "'%s' : %s" % (path, msg)) diff --git a/tools/livecd-creator b/tools/livecd-creator index b401190..137aecd 100755 --- a/tools/livecd-creator +++ b/tools/livecd-creator @@ -36,7 +36,7 @@ def parse_options(args): imgopt = optparse.OptionGroup(parser, "Image options", "These options define the created image.") imgopt.add_option("-c", "--config", type="string", dest="kscfg", - help="Path to kickstart config file") + help="Path or url to kickstart config file") imgopt.add_option("-b", "--base-on", type="string", dest="base_on", help="Add packages to an existing live CD iso9660 image.") imgopt.add_option("-f", "--fslabel", type="string", dest="fs_label", @@ -67,7 +67,7 @@ def parse_options(args): help=optparse.SUPPRESS_HELP) (options, args) = parser.parse_args() - if not options.kscfg or not os.path.isfile(options.kscfg): + if not options.kscfg or not imgcreate.verify_kickstart(options.kscfg): raise Usage("Kickstart config '%s' does not exist" %(options.kscfg,)) if options.base_on and not os.path.isfile(options.base_on): raise Usage("Live CD ISO '%s' does not exist" %(options.base_on,)) -- 1.5.5.2 From maillist at diffingo.com Fri Oct 17 02:44:38 2008 From: maillist at diffingo.com (Stewart Adam) Date: Thu, 16 Oct 2008 22:44:38 -0400 Subject: [Fedora-livecd-list] Using both EFI and MBR at once on Live USB images Message-ID: <48F7FC16.9090201@diffingo.com> Hi, I've been trying to get the Fedora 10 Snapshot ISO to work boot via a USB key on my MacBook Pro using the --mactel option, and while I think GRUB's EFI image has a bug in it since no kernels boot [1], it gave me another idea. I thought it would be pretty cool if we could synchronize the GPT and MSDOS partition tables (complete with MBR), similar to how it's used in BootCamp/rEFIt so that any given live media could be used with EFI or with BIOS. I've packaged rEFIt [2] (which includes gptsync) but since there's only one partition on the Live USB media, gptsync immediately says "nothing to do" and exits. One way around this would be to use two-partitions on the live media, but that would troublesome since Windows will only recognize the first readable partition. The other way would be to patch gptsync and have it translate the EFI partition to "FAT32" for the MSDOS partition table. I'd like to hear your comments/feedback on this - Do you think it's a feature worth spending time on? Any other ideas for workarounds? Stewart [1] https://bugzilla.redhat.com/show_bug.cgi?id=466954 [2] http://downloads.diffingo.com/refit-0.11-1.fc9.src.rpm From katzj at redhat.com Fri Oct 17 03:25:14 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 16 Oct 2008 23:25:14 -0400 Subject: [Fedora-livecd-list] Using both EFI and MBR at once on Live USB images In-Reply-To: <48F7FC16.9090201@diffingo.com> References: <48F7FC16.9090201@diffingo.com> Message-ID: <1224213914.370.85.camel@aglarond.local> On Thu, 2008-10-16 at 22:44 -0400, Stewart Adam wrote: > I thought it would be pretty cool if we could synchronize the GPT and > MSDOS partition tables (complete with MBR), similar to how it's used in > BootCamp/rEFIt so that any given live media could be used with EFI or with > BIOS. I've packaged rEFIt [2] (which includes gptsync) but since there's > only one partition on the Live USB media, gptsync immediately says "nothing > to do" and exits. One way around this would be to use two-partitions on the > live media, but that would troublesome since Windows will only recognize the > first readable partition. The other way would be to patch gptsync and have > it translate the EFI partition to "FAT32" for the MSDOS partition table. First for the obvious question -- did you change the partition table on the USB stick you're testing with to be a GPT partition table? And then had one partition inside that? You can't change the partition type of the GPT partition to anything other than 'ee' in the MSDOS partition table if you want to keep EFI happy with it. Multiple partitions might be a workaround, although, yeah, it would kind of suck. Probably worth trying I guess. Jeremy From maillist at diffingo.com Fri Oct 17 12:00:37 2008 From: maillist at diffingo.com (Stewart Adam) Date: Fri, 17 Oct 2008 08:00:37 -0400 Subject: [Fedora-livecd-list] Using both EFI and MBR at once on Live USB images In-Reply-To: <1224213914.370.85.camel@aglarond.local> References: <48F7FC16.9090201@diffingo.com> <1224213914.370.85.camel@aglarond.local> Message-ID: <48F87E65.6040104@diffingo.com> Jeremy Katz wrote: > First for the obvious question -- did you change the partition table on > the USB stick you're testing with to be a GPT partition table? And then > had one partition inside that? > > You can't change the partition type of the GPT partition to anything > other than 'ee' in the MSDOS partition table if you want to keep EFI > happy with it. > > Multiple partitions might be a workaround, although, yeah, it would kind > of suck. Probably worth trying I guess. > > Jeremy > > -- > Fedora-livecd-list mailing list > Fedora-livecd-list at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-livecd-list For the tests I just created a new stick with --mactel (it told me it would erase all data and make a GPT table for me) and from there I just ran gptsync which said "nothing to do". So essentially, it's still an unmodified USB stick... Stewart From katzj at fedoraproject.org Fri Oct 17 13:10:30 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Fri, 17 Oct 2008 13:10:30 +0000 (UTC) Subject: [Fedora-livecd-list] 4 commits - imgcreate/kickstart.py tools/livecd-iso-to-disk.sh Message-ID: <20081017131030.874ACC046B@lists.fedorahosted.org> imgcreate/kickstart.py | 19 +++++++++++-------- tools/livecd-iso-to-disk.sh | 40 +++++++++++++++++++++++++++++++--------- 2 files changed, 42 insertions(+), 17 deletions(-) New commits: commit 3e8101049b2adfee2898bdf633f58671c792dd60 Author: Jeremy Katz Date: Fri Oct 17 09:08:36 2008 -0400 Fix setting up persistent home on mtd for XO diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index 0475b0a..0a1144d 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -513,7 +513,7 @@ fi if [ -n "$xo" ]; then echo "Setting up /boot/olpc.fth file" args=$(egrep "^[ ]*append" $USBMNT/$SYSLINUXPATH/isolinux.cfg |head -n1 |sed -e 's/.*initrd=[^ ]*//') - if [ -n "$xonohome" -a ! -f $USBMNT/LiveOS/$HOMEFILE ]; then + if [ -z "$xonohome" -a ! -f $USBMNT/LiveOS/$HOMEFILE ]; then args="$args persistenthome=mtd0" fi if [ ! -d $USBMNT/boot ]; then mkdir -p $USBMNT/boot ; fi commit 37a6ca3b713d23ca0da39eb501b83aab42a202e3 Author: Jeremy Katz Date: Fri Oct 17 09:08:20 2008 -0400 More integer vs string munging diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index 195786e..0475b0a 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -228,6 +228,7 @@ cryptedhome=1 keephome=1 homesizemb=0 swapsizemb=0 +overlaysizemb=0 HOMEFILE="home.img" while [ $# -gt 2 ]; do @@ -322,21 +323,21 @@ elif [ -n "$mactel" ]; then fi -if [ -n "$overlaysizemb" -a "$USBFS" = "vfat" ]; then +if [ "$overlaysizemb" -gt 0 -a "$USBFS" = "vfat" ]; then if [ "$overlaysizemb" -gt 2047 ]; then echo "Can't have an overlay of 2048MB or greater on VFAT" exitclean fi fi -if [ -n "$homesizemb" -a "$USBFS" = "vfat" ]; then +if [ "$homesizemb" -gt 0 -a "$USBFS" = "vfat" ]; then if [ "$homesizemb" -gt 2047 ]; then echo "Can't have a home overlay greater than 2048MB on VFAT" exitclean fi fi -if [ -n "$swapsizemb" -a "$USBFS" = "vfat" ]; then +if [ "$swapsizemb" -gt 0 -a "$USBFS" = "vfat" ]; then if [ "$swapsizemb" -gt 2047 ]; then echo "Can't have a swap file greater than 2048MB on VFAT" exitclean @@ -376,7 +377,7 @@ free=$(df -B1M $USBDEV |tail -n 1 |awk {'print $4;'}) if [ $(($overlaysizemb + $homesizemb + $livesize + $swapsizemb)) -gt $(($free + $tbd)) ]; then echo "Unable to fit live image + overlay on available space on USB stick" echo "Size of live image: $livesize" - [ -n "$overlaysizemb" ] && echo "Overlay size: $overlaysizemb" + [ "$overlaysizemb" -gt 0 ] && echo "Overlay size: $overlaysizemb" [ "$homesizemb" -gt 0 ] && echo "Home overlay size: $homesizemb" [ "$swapsizemb" -gt 0 ] && echo "Home overlay size: $swapsizemb" echo "Available space: $(($free + $tbd))" @@ -461,7 +462,7 @@ echo "Updating boot config file" sed -i -e "s/CDLABEL=[^ ]*/$USBLABEL/" -e "s/rootfstype=[^ ]*/rootfstype=$USBFS/" $BOOTCONFIG if [ -n "$kernelargs" ]; then sed -i -e "s/liveimg/liveimg ${kernelargs}/" $BOOTCONFIG ; fi -if [ -n "$overlaysizemb" ]; then +if [ "$overlaysizemb" -gt 0 ]; then echo "Initializing persistent overlay file" OVERFILE="overlay-$( /lib/udev/vol_id -l $USBDEV )-$( /lib/udev/vol_id -u $USBDEV )" if [ "$USBFS" = "vfat" ]; then commit 8b11d8f99119c4ff5f922286fbec67d8406b09a0 Author: Orion Poplawski Date: Fri Oct 17 09:07:13 2008 -0400 Support --defaultdesktop livecd-creator ignores xconfig --defaultdesktop parameter. I have: xconfig --startxonboot --defaultdesktop=KDE but no /etc/sysconfig/desktop file is written. This works for me, and seems pretty comparable to the code in anaconda. diff --git a/imgcreate/kickstart.py b/imgcreate/kickstart.py index d911925..01359e7 100644 --- a/imgcreate/kickstart.py +++ b/imgcreate/kickstart.py @@ -226,14 +226,17 @@ class ServicesConfig(KickstartConfig): class XConfig(KickstartConfig): """A class to apply a kickstart X configuration to a system.""" def apply(self, ksxconfig): - if not ksxconfig.startX: - return - f = open(self.path("/etc/inittab"), "rw+") - buf = f.read() - buf = buf.replace("id:3:initdefault", "id:5:initdefault") - f.seek(0) - f.write(buf) - f.close() + if ksxconfig.startX: + f = open(self.path("/etc/inittab"), "rw+") + buf = f.read() + buf = buf.replace("id:3:initdefault", "id:5:initdefault") + f.seek(0) + f.write(buf) + f.close() + if ksxconfig.defaultdesktop: + f = open(self.path("/etc/sysconfig/desktop"), "w") + f.write("DESKTOP="+ksxconfig.defaultdesktop+"\n") + f.close() class RPMMacroConfig(KickstartConfig): """A class to apply the specified rpm macros to the filesystem""" commit 0aadbdc442191e395ee7579716035df55c8d3ded Author: Jeremy Katz Date: Wed Oct 15 13:36:38 2008 -0400 Add support for setting up a swap Allow setting up a swap on the usb stick/sd card. Also includes a fix from David Zuhn to use ints instead of depending on empty strings so we don't get an error from the bash math diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index d48f922..195786e 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -226,6 +226,9 @@ fi cryptedhome=1 keephome=1 +homesizemb=0 +swapsizemb=0 + HOMEFILE="home.img" while [ $# -gt 2 ]; do case $1 in @@ -237,6 +240,10 @@ while [ $# -gt 2 ]; do homesizemb=$2 shift ;; + --swap-size-mb) + swapsizemb=$2 + shift + ;; --crypted-home) cryptedhome=1 ;; @@ -329,6 +336,13 @@ if [ -n "$homesizemb" -a "$USBFS" = "vfat" ]; then fi fi +if [ -n "$swapsizemb" -a "$USBFS" = "vfat" ]; then + if [ "$swapsizemb" -gt 2047 ]; then + echo "Can't have a swap file greater than 2048MB on VFAT" + exitclean + fi +fi + # FIXME: would be better if we had better mountpoints CDMNT=$(mktemp -d /media/cdtmp.XXXXXX) mount -o loop,ro "$ISO" $CDMNT || exitclean @@ -337,7 +351,7 @@ mount $mountopts $USBDEV $USBMNT || exitclean trap exitclean SIGINT SIGTERM -if [ -f "$USBMNT/LiveOS/$HOMEFILE" -a -n "$keephome" -a -n "$homesizemb" ]; then +if [ -f "$USBMNT/LiveOS/$HOMEFILE" -a -n "$keephome" -a "$homesizemb" -gt 0 ]; then echo "ERROR: Requested keeping existing /home and specified a size for /home" echo "Please either don't specify a size or specify --delete-home" exitclean @@ -359,11 +373,12 @@ fi livesize=$(du -s -B 1M $check | awk {'print $1;'}) free=$(df -B1M $USBDEV |tail -n 1 |awk {'print $4;'}) -if [ $(($overlaysizemb + $homesizemb + $livesize)) -gt $(($free + $tbd)) ]; then +if [ $(($overlaysizemb + $homesizemb + $livesize + $swapsizemb)) -gt $(($free + $tbd)) ]; then echo "Unable to fit live image + overlay on available space on USB stick" echo "Size of live image: $livesize" [ -n "$overlaysizemb" ] && echo "Overlay size: $overlaysizemb" - [ -n "$homesizemb" ] && echo "Home overlay size: $homesizemb" + [ "$homesizemb" -gt 0 ] && echo "Home overlay size: $homesizemb" + [ "$swapsizemb" -gt 0 ] && echo "Home overlay size: $swapsizemb" echo "Available space: $(($free + $tbd))" exitclean fi @@ -459,7 +474,13 @@ if [ -n "$overlaysizemb" ]; then sed -i -e "s/\ ro\ /\ rw\ /" $BOOTCONFIG fi -if [ -n "$homesizemb" ]; then +if [ "$swapsizemb" -gt 0 ]; then + echo "Initializing swap file" + dd if=/dev/zero of=$USBMNT/LiveOS/swap.img count=$swapsizemb bs=1M + mkswap -f $USBMNT/LiveOS/swap.img +fi + +if [ "$homesizemb" -gt 0 ]; then echo "Initializing persistent /home" if [ "$USBFS" = "vfat" ]; then # vfat can't handle sparse files From katzj at redhat.com Fri Oct 17 13:50:19 2008 From: katzj at redhat.com (Jeremy Katz) Date: Fri, 17 Oct 2008 09:50:19 -0400 Subject: [Fedora-livecd-list] Using both EFI and MBR at once on Live USB images In-Reply-To: <48F87E65.6040104@diffingo.com> References: <48F7FC16.9090201@diffingo.com> <1224213914.370.85.camel@aglarond.local> <48F87E65.6040104@diffingo.com> Message-ID: <1224251419.13356.46.camel@aglarond.local> On Fri, 2008-10-17 at 08:00 -0400, Stewart Adam wrote: > Jeremy Katz wrote: > > First for the obvious question -- did you change the partition table on > > the USB stick you're testing with to be a GPT partition table? And then > > had one partition inside that? > > > > You can't change the partition type of the GPT partition to anything > > other than 'ee' in the MSDOS partition table if you want to keep EFI > > happy with it. > > > > Multiple partitions might be a workaround, although, yeah, it would kind > > of suck. Probably worth trying I guess. > > For the tests I just created a new stick with --mactel (it told me > it would erase all data and make a GPT table for me) and from there > I just ran gptsync which said "nothing to do". So essentially, it's > still an unmodified USB stick... With a quick look at the gptsync code, this might actually be a bug in gptsync -- it looks like it only does the sync if the number of partitions doesn't match. So getting rid of that check and making it unconditionally sync might help Jeremy From katzj at redhat.com Fri Oct 17 13:53:05 2008 From: katzj at redhat.com (Jeremy Katz) Date: Fri, 17 Oct 2008 09:53:05 -0400 Subject: [Fedora-livecd-list] [PATCH] Allow for http:// or ftp:// urls to be passed in to the image creators In-Reply-To: <1224191503-7562-1-git-send-email-bkearney@redhat.com> References: <1224191503-7562-1-git-send-email-bkearney@redhat.com> Message-ID: <1224251585.13356.49.camel@aglarond.local> On Thu, 2008-10-16 at 17:11 -0400, Bryan Kearney wrote: > @@ -22,6 +22,7 @@ import shutil > import subprocess > import time > import logging > +import urllib I'd rather see urlgrabber used than raw urllib. The API is nicer and it will get us more functionality in the long run > @@ -34,6 +35,19 @@ import pykickstart.version as ksversion > import imgcreate.errors as errors > import imgcreate.fs as fs > > +def path_is_url(path): > + pupper = path.upper() > + return (pupper.startswith("HTTP://")) or (pupper.startswith("FTP://")) > + > +def verify_kickstart(path): > + verified = False > + if not path is None: > + if path_is_url(path): > + verified = True > + else: > + verified = os.path.isfile(path) > + return verified > + Given that this doesn't cover all of the URL types that are supported (https:// and file:// for example), it's probably better to just assume that it's "valid" and ensure that we have good error handling for when its not. Especially since we'll have to have that error handling anyway for the case of an improper URL being given. Jeremy From bkearney at redhat.com Fri Oct 17 14:22:44 2008 From: bkearney at redhat.com (Bryan Kearney) Date: Fri, 17 Oct 2008 10:22:44 -0400 Subject: [Fedora-livecd-list] Resending the remote kickstart patch Message-ID: <1224253365-17158-1-git-send-email-bkearney@redhat.com> This is a resend of the patch. Changes include: - Requireing and using python-urlgrabber instead of urllib - Allowing for file:// based urls. - Any string which starts with http://, ftp://, or file:// is assumed valid and will be caught during parsing. - Rebased with latest upstream changes. -- bk From bkearney at redhat.com Fri Oct 17 14:22:45 2008 From: bkearney at redhat.com (Bryan Kearney) Date: Fri, 17 Oct 2008 10:22:45 -0400 Subject: [Fedora-livecd-list] [PATCH] Allow for http://, ftp://, or file:// urls to be passed in to the image creators using urlgrabber. In-Reply-To: <1224253365-17158-1-git-send-email-bkearney@redhat.com> References: <1224253365-17158-1-git-send-email-bkearney@redhat.com> Message-ID: <1224253365-17158-2-git-send-email-bkearney@redhat.com> From: root --- imgcreate/__init__.py | 2 ++ imgcreate/kickstart.py | 22 +++++++++++++++++++++- livecd-tools.spec | 1 + tools/livecd-creator | 4 ++-- 4 files changed, 26 insertions(+), 3 deletions(-) diff --git a/imgcreate/__init__.py b/imgcreate/__init__.py index bcfb40e..7cf30b8 100644 --- a/imgcreate/__init__.py +++ b/imgcreate/__init__.py @@ -34,6 +34,7 @@ Also exported are: - CreatorError - all exceptions throw are of this type - FSLABEL_MAXLEN - the length to which LoopImageCreator.fslabel is truncated - read_kickstart() - a utility function for kickstart parsing + - verify_kickstart() - a utility function to verify the existence of a kickstart file - build_name() - a utility to construct an image name Each of the creator classes are designed to be subclassable, allowing the user @@ -63,6 +64,7 @@ __all__ = ( 'LoopImageCreator', 'FSLABEL_MAXLEN', 'read_kickstart', + 'verify_kickstart', 'construct_name', 'setup_logging', ) diff --git a/imgcreate/kickstart.py b/imgcreate/kickstart.py index 01359e7..a6993ad 100644 --- a/imgcreate/kickstart.py +++ b/imgcreate/kickstart.py @@ -22,6 +22,7 @@ import shutil import subprocess import time import logging +import urlgrabber import rhpl.keyboard @@ -34,6 +35,21 @@ import pykickstart.version as ksversion import imgcreate.errors as errors import imgcreate.fs as fs +def path_is_url(path): + pupper = path.upper() + return (pupper.startswith("HTTP://")) \ + or (pupper.startswith("FTP://")) \ + or (pupper.startswith("FILE://")) + +def verify_kickstart(path): + verified = False + if not path is None: + if path_is_url(path): + verified = True + else: + verified = os.path.isfile(path) + return verified + def read_kickstart(path): """Parse a kickstart file and return a KickstartParser instance. @@ -47,7 +63,11 @@ def read_kickstart(path): version = ksversion.makeVersion() ks = ksparser.KickstartParser(version) try: - ks.readKickstart(path) + if path_is_url(path): + ksfile = urlgrabber.urlgrab(path) + ks.readKickstart(ksfile) + else: + ks.readKickstart(path) except IOError, (err, msg): raise errors.KickstartError("Failed to read kickstart file " "'%s' : %s" % (path, msg)) diff --git a/livecd-tools.spec b/livecd-tools.spec index 681f6cf..99c6d99 100644 --- a/livecd-tools.spec +++ b/livecd-tools.spec @@ -21,6 +21,7 @@ Requires: pykickstart >= 0.96 Requires: dosfstools >= 2.11-8 Requires: isomd5sum Requires: rhpl +Requires: python-urlgrabber %ifarch %{ix86} x86_64 Requires: syslinux %endif diff --git a/tools/livecd-creator b/tools/livecd-creator index b401190..137aecd 100755 --- a/tools/livecd-creator +++ b/tools/livecd-creator @@ -36,7 +36,7 @@ def parse_options(args): imgopt = optparse.OptionGroup(parser, "Image options", "These options define the created image.") imgopt.add_option("-c", "--config", type="string", dest="kscfg", - help="Path to kickstart config file") + help="Path or url to kickstart config file") imgopt.add_option("-b", "--base-on", type="string", dest="base_on", help="Add packages to an existing live CD iso9660 image.") imgopt.add_option("-f", "--fslabel", type="string", dest="fs_label", @@ -67,7 +67,7 @@ def parse_options(args): help=optparse.SUPPRESS_HELP) (options, args) = parser.parse_args() - if not options.kscfg or not os.path.isfile(options.kscfg): + if not options.kscfg or not imgcreate.verify_kickstart(options.kscfg): raise Usage("Kickstart config '%s' does not exist" %(options.kscfg,)) if options.base_on and not os.path.isfile(options.base_on): raise Usage("Live CD ISO '%s' does not exist" %(options.base_on,)) -- 1.5.5.2 From maillist at diffingo.com Fri Oct 17 14:52:06 2008 From: maillist at diffingo.com (Stewart Adam) Date: Fri, 17 Oct 2008 10:52:06 -0400 Subject: [Fedora-livecd-list] Using both EFI and MBR at once on Live USB images In-Reply-To: <1224251419.13356.46.camel@aglarond.local> References: <48F7FC16.9090201@diffingo.com> <1224213914.370.85.camel@aglarond.local> <48F87E65.6040104@diffingo.com> <1224251419.13356.46.camel@aglarond.local> Message-ID: <48F8A696.8020601@diffingo.com> Jeremy Katz wrote: > On Fri, 2008-10-17 at 08:00 -0400, Stewart Adam wrote: >> Jeremy Katz wrote: >>> First for the obvious question -- did you change the partition table on >>> the USB stick you're testing with to be a GPT partition table? And then >>> had one partition inside that? >>> >>> You can't change the partition type of the GPT partition to anything >>> other than 'ee' in the MSDOS partition table if you want to keep EFI >>> happy with it. >>> >>> Multiple partitions might be a workaround, although, yeah, it would kind >>> of suck. Probably worth trying I guess. >> For the tests I just created a new stick with --mactel (it told me >> it would erase all data and make a GPT table for me) and from there >> I just ran gptsync which said "nothing to do". So essentially, it's >> still an unmodified USB stick... > > With a quick look at the gptsync code, this might actually be a bug in > gptsync -- it looks like it only does the sync if the number of > partitions doesn't match. So getting rid of that check and making it > unconditionally sync might help > > Jeremy > > Hm, I tried again and I got my errors mixed up. It actually says that there's no data partition (only the EFI system one) so there's no need to sync as the EFI system partition is always marked EFI Protective in the MBR. If it isn't, then installing anything to the MBR risks erasing the entire GPT data structure... So the after all the only option seems to be to use multiple partitions, which I don't think will work well. I'm still going to try a few things though... If GRUB can install to the MBR on the "EFI Protective" partition and it realizes that "EFI Protective" is just FAT32, things might work out on a single partition. Stewart From katzj at redhat.com Fri Oct 17 16:19:13 2008 From: katzj at redhat.com (Jeremy Katz) Date: Fri, 17 Oct 2008 12:19:13 -0400 Subject: [Fedora-livecd-list] Resending the remote kickstart patch In-Reply-To: <1224253365-17158-1-git-send-email-bkearney@redhat.com> References: <1224253365-17158-1-git-send-email-bkearney@redhat.com> Message-ID: <1224260353.13356.51.camel@aglarond.local> On Fri, 2008-10-17 at 10:22 -0400, Bryan Kearney wrote: > - Any string which starts with http://, ftp://, or file:// is assumed valid and will be caught during parsing. You didn't seem to entirely follow this part -- I'm not convinced that trying to do any validation really matters here. Instead, as long as we got _something_ as the kickstart config, we can accept it and then just have to nicely catch errors when retrieving it. Which would then mean not needing the verify_kickstart() and path_is_url methods (especially as there are cases they don't handle) Jeremy From todd_cooper at yahoo.com Mon Oct 20 16:57:21 2008 From: todd_cooper at yahoo.com (Todd Cooper) Date: Mon, 20 Oct 2008 09:57:21 -0700 (PDT) Subject: [Fedora-livecd-list] xconfig does not work Message-ID: <103101.35336.qm@web62402.mail.re1.yahoo.com> Still have the problem and I can get into the non-gui login (Shift-Ctrl-F1) /var/log/messages and /var/log/gdm/* have no inkling to what is wrong. I do not see any ~/.xsession-errors. Any other place to look for errors? ----- Original Message ---- From: Jeremy Katz To: Todd Cooper ; fedora-livecd-list at redhat.com Sent: Thursday, October 2, 2008 10:39:35 AM Subject: Re: [Fedora-livecd-list] xconfig does not work On Wed, 2008-10-01 at 09:13 -0700, Todd Cooper wrote: > I added small things to the desktop kickstart (attached) and the > xconfig no longer works. > The biggest change is adding mysql to the rpms to be added. > The .ks file attached has xconfig commented out. > > When I boot it comes up with the X login and when I click on the login > user, it starts the login and then goes back to the login screen. > Even if I login as root it does the same thing. You probably are missing whatever the xsession is trying to run. If you switch to a tty and check ~/.xsession-errors, you'll probably see something more enlightening Jeremy -- Fedora-livecd-list mailing list Fedora-livecd-list at redhat.com https://www.redhat.com/mailman/listinfo/fedora-livecd-list -------------- next part -------------- An HTML attachment was scrubbed... URL: From khushi.pk at gmail.com Mon Oct 20 19:50:41 2008 From: khushi.pk at gmail.com (Khushbu Mohta) Date: Tue, 21 Oct 2008 01:20:41 +0530 Subject: [Fedora-livecd-list] Kobold Project Message-ID: <56cf9fde0810201250q3065b36ftdc6e0b22eec7497@mail.gmail.com> Hi Everyone, As part of the Summer Training[1] organised by Kushal Das[2] and other Durgapur LUG[3] Team-members at IRC channel #dgplug, I along with my colleague Prakhar Agarwal have worked on Kobold Project [4]. We were suggested this project by Rahul Sundaram(mether). *About Kobold:* Kobold is minimalistic tool for automated Fedora Live CD installation on USB stick in a sequential manner. It uses HAL and D-Bus to discover USB stick plugging and livecd-iso-to-disk for the actual installation process. * Some pitfalls*: (1) It was sequential. (2) It could only work on a Single USB stick at a time with hassles of plugging-in and plugging-out the USB sticks if more than one USB stick need to be used. (3) The USB sticks had to be plugged-in after initiating the script. The script was not able to detect already plugged-in USB sticks. *Our Tweaks:* First of all, we eradicated the need to plug-in the USB sticks after initiating the script. That is, we added the functionality of detecting already plugged-in USB storage devices. Secondly, the installation process is changed from sequential to parallel. We have incorporated threading concept by using "threading" module to enable installation process on multiple USB sticks simultaneously. Code is available at http://dgplug.org/training/trunk/khushbu/Kobold/kobold3.py *Usability/Benefits*: The Install-fests will be much more fun. Just plug-in all the USB storage devices, run the script and you will have Fedora installed in all of them within no time!! No need to wait for the installation process to get over with one stick and then plug-in another. *References:* [1] http://wiki.dgplug.org/index.php/SummerTraining [2] http://kushaldas.in [3] http://dgplug.org [4] http://trac2.assembla.com/kobold-project We hope that we have been able to contribute in some way. We wish to add this code to LiveUSB-Creator(which works on a single USB drive right now) and for that we need your comments/feedback. Thank You! -- Khushbu Mohta khushi.pk at gmail.com -- Khushbu Mohta khushi.pk at gmail.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From opensource at till.name Mon Oct 20 20:18:00 2008 From: opensource at till.name (Till Maas) Date: Mon, 20 Oct 2008 22:18:00 +0200 Subject: [Fedora-livecd-list] RFE: implant locale information into isos like anaconda implants md5sums Message-ID: <200810202218.04081.opensource@till.name> Hiyas, I have an idea to make it easier for people to create an iso image with the desired locale and keyboard layout for local distribution. The anaconda project created a tool "isomd5sum" that somehow implants the md5sum of major parts of the iso into the iso file itself. I did not yet find the time to investigate how they do it. I would find it very useful, if the same technique could be used to store the default locale and keyboard layout within the iso image for live cds. What do you think about this idea? Regards, Till -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 827 bytes Desc: This is a digitally signed message part. URL: From katzj at redhat.com Mon Oct 20 20:38:04 2008 From: katzj at redhat.com (Jeremy Katz) Date: Mon, 20 Oct 2008 16:38:04 -0400 Subject: [Fedora-livecd-list] RFE: implant locale information into isos like anaconda implants md5sums In-Reply-To: <200810202218.04081.opensource@till.name> References: <200810202218.04081.opensource@till.name> Message-ID: <1224535084.13204.34.camel@aglarond.local> On Mon, 2008-10-20 at 22:18 +0200, Till Maas wrote: > I have an idea to make it easier for people to create an iso image with the > desired locale and keyboard layout for local distribution. > > The anaconda project created a tool "isomd5sum" that somehow implants the > md5sum of major parts of the iso into the iso file itself. I did not yet find > the time to investigate how they do it. I would find it very useful, if the > same technique could be used to store the default locale and keyboard layout > within the iso image for live cds. > > What do you think about this idea? The md5sum embedding is done by putting the information into the application specific area of the ISO. The problem is that this is a (very) limited space and we're already using almost all of it to store md5sums. Also, I think (hope? :) that people are using the live images as much or more off of USB sticks and so I'd rather not make any choices like this which aren't generally usable no matter what type of medium you're running the live image from. Jeremy From kwizart at gmail.com Mon Oct 20 22:04:15 2008 From: kwizart at gmail.com (KH KH) Date: Tue, 21 Oct 2008 00:04:15 +0200 Subject: [Fedora-livecd-list] RFE: implant locale information into isos like anaconda implants md5sums In-Reply-To: <200810202218.04081.opensource@till.name> References: <200810202218.04081.opensource@till.name> Message-ID: 2008/10/20 Till Maas : > Hiyas, > > I have an idea to make it easier for people to create an iso image with the > desired locale and keyboard layout for local distribution. > > The anaconda project created a tool "isomd5sum" that somehow implants the > md5sum of major parts of the iso into the iso file itself. I did not yet find > the time to investigate how they do it. I would find it very useful, if the > same technique could be used to store the default locale and keyboard layout > within the iso image for live cds. > > What do you think about this idea? Hello ! The "whole" problematic of localized LiveCD isn't only that language, keyboard, date (but also firefox/thunderbird dictionaries) are thoses from a given locale option. But also that translation texts (/usr/share/locale/* ) are also removed to only keep the desired locale. This will save around 200Mo from a 700Mo international spin. (tested in the fr_FR LiveCD case). So maybe it could be possible to store locales in a special place so it can be removed easily from the international spin when the iso is getting transferered to the (LiveUSB) device. But for now i'm still using (from the fedora-livecd-desktop-fr_FR.ks ): %pre # Only install the needed locales mkdir -p /etc/rpm echo "%_install_langs en:fr:fr_FR:fr-FR" > /etc/rpm/macros.lang Nicolas (kwizart) From katzj at redhat.com Tue Oct 21 01:09:13 2008 From: katzj at redhat.com (Jeremy Katz) Date: Mon, 20 Oct 2008 21:09:13 -0400 Subject: [Fedora-livecd-list] RFE: implant locale information into isos like anaconda implants md5sums In-Reply-To: References: <200810202218.04081.opensource@till.name> Message-ID: <1224551353.13204.182.camel@aglarond.local> On Tue, 2008-10-21 at 00:04 +0200, KH KH wrote: > The "whole" problematic of localized LiveCD isn't only that language, > keyboard, date (but also firefox/thunderbird dictionaries) are thoses > from a given locale option. > But also that translation texts (/usr/share/locale/* ) are also > removed to only keep the desired locale. There aren't any translations removed from the default desktop live image. I can't speak to what's done for the other live images > This will save around 200Mo > from a 700Mo international spin. (tested in the fr_FR LiveCD case). > So maybe it could be possible to store locales in a special place so > it can be removed easily from the international spin when the iso is > getting transferered to the (LiveUSB) device. The problems with removing languages from being installed start with a) No easy way for someone to go back and add a locale after the fact b) Not having bits installed means that if you're using deltarpms, you can't really take advantage of deltas > But for now i'm still using (from the fedora-livecd-desktop-fr_FR.ks ): > %pre > # Only install the needed locales > mkdir -p /etc/rpm > echo "%_install_langs en:fr:fr_FR:fr-FR" > /etc/rpm/macros.lang You do realize that %pre doesn't even run at all with livecd-creator? Jeremy From opensource at till.name Tue Oct 21 08:45:16 2008 From: opensource at till.name (Till Maas) Date: Tue, 21 Oct 2008 10:45:16 +0200 Subject: [Fedora-livecd-list] RFE: implant locale information into isos like anaconda implants md5sums In-Reply-To: <1224535084.13204.34.camel@aglarond.local> References: <200810202218.04081.opensource@till.name> <1224535084.13204.34.camel@aglarond.local> Message-ID: <200810211045.17789.opensource@till.name> On Mon October 20 2008, Jeremy Katz wrote: > The md5sum embedding is done by putting the information into the > application specific area of the ISO. The problem is that this is a > (very) limited space and we're already using almost all of it to store > md5sums. :-( > Also, I think (hope? :) that people are using the live images as much or > more off of USB sticks and so I'd rather not make any choices like this > which aren't generally usable no matter what type of medium you're > running the live image from. Currently I am more often booting a new iso image in qemu-kvm, than one from a USB stick. There it would be nice to simply run some "implant-locale" script to the iso, without the need to copy all contents onto a virtual USB device. And the information from the iso could also be used by a livecd-iso-to-disk script that can set the keyboard layout and locale somehow. Regards, Till -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 827 bytes Desc: This is a digitally signed message part. URL: From muayyad.alsadi at ojuba.org Wed Oct 22 19:13:02 2008 From: muayyad.alsadi at ojuba.org (=?UTF-8?Q?=D9=85=D8=A4=D9=8A=D8=AF_=D8=A7=D9=84=D8=B3=D8=B9=D8=AF=D9=8A?=) Date: Wed, 22 Oct 2008 22:13:02 +0300 Subject: [Fedora-livecd-list] RFE: implant locale information into isos like anaconda implants md5sums In-Reply-To: <200810202218.04081.opensource@till.name> References: <200810202218.04081.opensource@till.name> Message-ID: <25788f0800db7f971099c8ec8a4ecd87@localhost> Hiyas you too, I love to see away to set the default language set at compose time in the ks file and anaconda should read the language and focus/highlight it by default on language selection dialog [in liveinst] my way of doing this is that I noticed that all initscripts source /etc/init.d/functions which sources /etc/profile.d/lang.sh which selects the language from /etc/sysonfig/i18n and other places so I added another place to that file which is the boot parameters https://bugzilla.redhat.com/show_bug.cgi?id=454105 and on our fedora-based Linux (www.Ojuba.org) the boot menu passes LANG=ar_JO.UTF-8 I consider this to be very useful to implement even if it's not the best way for example as a fail-safe language setting procedure or as a feature that will help us when gfxboot is upstreamed so that language menu implementation would be trivial but even if this is done, I think the right place for default language to be set is in .ks file and anaconda should respect it and highlight it by default [not skip the language selection completely] I tried to ask how todo that in fedora-buildsys-list https://www.redhat.com/archives/fedora-buildsys-list/2008-October/msg00000.html but everybody ignored me I think that anaconda already have such feature, as I noticed a line for language if self.listStore.get_value(iter, 1) == current: and another for the keyboard layour if self.modelStore.get_value(iter, 0) == default: so highlighting English is not hardcoded into anaconda, and it read current or default from a place I don't know we need to know what is that place and to let pungi or livecd-creator set the default there, that's it BTW: I suppose Hiyas is a greeting in your language, I would love to know which language is it and what does it mean O mankind! We created you from a single (pair) of a male and a female, and made you into nations and tribes, that ye may know each other (not that ye may despise (each other). Verily the most honoured of you in the sight of Allah is (he who is) the most righteous of you. And Allah has full knowledge and is well acquainted (with all things). -- meaning of V49:13 from Quran From sundaram at fedoraproject.org Wed Oct 22 22:01:15 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Thu, 23 Oct 2008 03:31:15 +0530 Subject: [Fedora-livecd-list] Re: [fedora-india] Kobold Project In-Reply-To: <56cf9fde0810221308g4237065ex2b3683e7ebf81167@mail.gmail.com> References: <56cf9fde0810201231x233ba850kce8d9c6213e130ef@mail.gmail.com> <48FEF95F.6010803@gmail.com> <56cf9fde0810221308g4237065ex2b3683e7ebf81167@mail.gmail.com> Message-ID: <48FFA2AB.5070709@fedoraproject.org> Khushbu Mohta wrote: > "Sankarshan (????????)" wrote: > > Hmm... so can this be used during FOSS.IN for the > LiveUSB station ? > > > Yes surely it can be used. We have removed all the possibles flaws and > tested it. Soon we will add the persistence overlay also. I believe it > will be a smooth ride in FOSS.IN One thing I would like to note is that, it is important we get this merged as part of some existing tool like livecd-creator or liveusb-creator. Independent scripts are great to prototype and get started but will rot if let as a isolated development. Jeremy Katz, Luke Macken or others? could offer some feedback on this. Rahul From prakhar at fedoraproject.org Thu Oct 23 00:11:44 2008 From: prakhar at fedoraproject.org (Prakhar Agarwal) Date: Thu, 23 Oct 2008 05:41:44 +0530 Subject: [Fedora-livecd-list] Re: [fedora-india] Kobold Project In-Reply-To: <48FFA2AB.5070709@fedoraproject.org> References: <56cf9fde0810201231x233ba850kce8d9c6213e130ef@mail.gmail.com> <48FEF95F.6010803@gmail.com> <56cf9fde0810221308g4237065ex2b3683e7ebf81167@mail.gmail.com> <48FFA2AB.5070709@fedoraproject.org> Message-ID: Rahul Sundaram wrote: > > One thing I would like to note is that, it is important we get this merged > as part of some existing tool like livecd-creator or liveusb-creator. > Independent scripts are great to prototype and get started but will rot if > let as a isolated development. > > Jeremy Katz, Luke Macken or others? could offer some feedback on this. > > +1 I believe that after some testing and enhancement, the code can find its way into some existing tool. Lets see what others feel about this. Cheers! -- Prakhar Agarwal Fedora Ambassador- Delhi, India -------------- next part -------------- An HTML attachment was scrubbed... URL: From orion at cora.nwra.com Thu Oct 23 21:49:33 2008 From: orion at cora.nwra.com (Orion Poplawski) Date: Thu, 23 Oct 2008 15:49:33 -0600 Subject: [Fedora-livecd-list] Why is /sbin not in %post path for livecd-creator? Message-ID: <4900F16D.2090305@cora.nwra.com> Why is /sbin not in %post path for livecd-creator? + chkconfig netfs off /tmp/ks-script-mzTh-j: line 32: chkconfig: command not found -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion at cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com From orion at cora.nwra.com Thu Oct 23 23:22:27 2008 From: orion at cora.nwra.com (Orion Poplawski) Date: Thu, 23 Oct 2008 17:22:27 -0600 Subject: [Fedora-livecd-list] Trouble mounting persistent home... Message-ID: <49010733.3080301@cora.nwra.com> I'm having trouble with my rawhide live usb devices and mounting the persistent home image. when the fedora-live init script runs I get: mount: /dev/sdb1 already mounted or /mnt/live busy mount doesn't show /dev/sdb1, but fuser -m /mnt/live lists basically every process starting with 1. /mnt/live: 1rce 2rc 3rc .... I wouldn't put it past issues with my setup, but I've tried to copy the %post from the fedora-release files pretty closely. Using latest git livecd tools. -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion at cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com From orion at cora.nwra.com Thu Oct 23 23:24:03 2008 From: orion at cora.nwra.com (Orion Poplawski) Date: Thu, 23 Oct 2008 17:24:03 -0600 Subject: [Fedora-livecd-list] Trouble mounting persistent home... In-Reply-To: <49010733.3080301@cora.nwra.com> References: <49010733.3080301@cora.nwra.com> Message-ID: <49010793.3090508@cora.nwra.com> Orion Poplawski wrote: > I'm having trouble with my rawhide live usb devices and mounting the > persistent home image. > > when the fedora-live init script runs I get: > > mount: /dev/sdb1 already mounted or /mnt/live busy > > mount doesn't show /dev/sdb1, but fuser -m /mnt/live lists basically > every process starting with 1. > > /mnt/live: 1rce 2rc 3rc .... > > I wouldn't put it past issues with my setup, but I've tried to copy the > %post from the fedora-release files pretty closely. Using latest git > livecd tools. > Oh, ls /mnt/live shows it empty at this point... -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion at cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com From orion at cora.nwra.com Thu Oct 23 23:32:33 2008 From: orion at cora.nwra.com (Orion Poplawski) Date: Thu, 23 Oct 2008 17:32:33 -0600 Subject: [Fedora-livecd-list] Trouble mounting persistent home... In-Reply-To: <49010793.3090508@cora.nwra.com> References: <49010733.3080301@cora.nwra.com> <49010793.3090508@cora.nwra.com> Message-ID: <49010991.5040406@cora.nwra.com> Orion Poplawski wrote: > Orion Poplawski wrote: >> I'm having trouble with my rawhide live usb devices and mounting the >> persistent home image. >> >> when the fedora-live init script runs I get: >> >> mount: /dev/sdb1 already mounted or /mnt/live busy >> >> mount doesn't show /dev/sdb1, but fuser -m /mnt/live lists basically >> every process starting with 1. >> >> /mnt/live: 1rce 2rc 3rc .... >> >> I wouldn't put it past issues with my setup, but I've tried to copy >> the %post from the fedora-release files pretty closely. Using latest >> git livecd tools. >> > > Oh, ls /mnt/live shows it empty at this point... > > Okay, fedora-live does: mount -o ro /dev/live /mnt/live if I do: mount /dev/live /mnt/live it works. Forget the fuser stuff, that's me being dumb. So, what's causing this? -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion at cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com From katzj at fedoraproject.org Mon Oct 27 22:16:51 2008 From: katzj at fedoraproject.org (Jeremy Katz) Date: Mon, 27 Oct 2008 22:16:51 +0000 (UTC) Subject: [Fedora-livecd-list] 4 commits - tools/livecd-iso-to-disk.sh Message-ID: <20081027221651.ED20A120156@lists.fedorahosted.org> tools/livecd-iso-to-disk.sh | 27 ++++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) New commits: commit 7cc4bfb25339ebc6d8cc233995f6e14a6fc6bc35 Author: Jeremy Katz Date: Mon Oct 27 17:15:16 2008 -0400 Minor fixes for mactel mode Make sure we use the new UUID of the created filesystme and remove the grub.conf to reduce confusion potential diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index ccbe694..497062f 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -133,6 +133,7 @@ createGPTLayout() { USBDEV=${device}1 /sbin/udevsettle /sbin/mkdosfs -n LIVE $USBDEV + USBLABEL="UUID=$(/lib/udev/vol_id -u $dev)" } checkGPT() { @@ -477,6 +478,7 @@ EOF # this is a little ugly, but it gets the "interesting" named config file BOOTCONFIG=$USBMNT/EFI/boot/boot?*.conf + rm -f $USBMNT/EFI/boot/grub.conf fi echo "Updating boot config file" commit c13700335b3296d8672c53ae599164ff0eb27934 Author: Jeremy Katz Date: Tue Oct 21 11:18:58 2008 -0400 Support copying the raw ext3fs to the usb stick For some cases, squashfs is pretty expensive and we have disk space (eg, on SD card for the XO). In those cases, we want to be able to ask for just the raw ext3fs.img to be copied. Make --xo default to doing so diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index b598fa4..ccbe694 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -274,10 +274,14 @@ while [ $# -gt 2 ]; do ;; --xo) xo=1 + skipcompress=1 ;; --xo-no-home) xonohome=1 ;; + --skipcompress) + skipcompress=1 + ;; --extra-kernel-args) kernelargs=$2 shift @@ -381,6 +385,11 @@ else tbd=0 fi livesize=$(du -s -B 1M $check | awk {'print $1;'}) +if [ -n "$skipcompress" ]; then + mount -o loop $CDMNT/LiveOS/squashfs.img $CDMNT + livesize=$(du -s -B 1M $CDMNT/LiveOS/ext3fs.img | awk {'print $1;'}) + umount $CDMNT +fi free=$(df -B1M $USBDEV |tail -n 1 |awk {'print $4;'}) if [ $(($overlaysizemb + $homesizemb + $livesize + $swapsizemb)) -gt $(($free + $tbd)) ]; then @@ -415,7 +424,11 @@ echo "Copying live image to USB stick" [ ! -d $USBMNT/LiveOS ] && mkdir $USBMNT/LiveOS [ -n "$keephome" -a -f "$USBMNT/$HOMEFILE" ] && mv $USBMNT/$HOMEFILE $USBMNT/LiveOS/$HOMEFILE # cases without /LiveOS are legacy detection, remove for F10 -if [ -f $CDMNT/LiveOS/squashfs.img ]; then +if [ -n "$skipcompress" -a -f $CDMNT/LiveOS/squashfs.img ]; then + mount -o loop $CDMNT/LiveOS/squashfs.img $CDMNT + cp $CDMNT/LiveOS/ext3fs.img $USBMNT/LiveOS/ext3fs.img || (umount $CDMNT ; exitclean) + umount $CDMNT +elif [ -f $CDMNT/LiveOS/squashfs.img ]; then cp $CDMNT/LiveOS/squashfs.img $USBMNT/LiveOS/squashfs.img || exitclean elif [ -f $CDMNT/squashfs.img ]; then cp $CDMNT/squashfs.img $USBMNT/LiveOS/squashfs.img || exitclean commit 0896e5bc135b1c7f14254dbaf0a97c9a0a6f89b8 Author: Jeremy Katz Date: Tue Oct 21 11:14:17 2008 -0400 Default to resetting the overlay on XO diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index c224230..b598fa4 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -525,6 +525,7 @@ if [ -n "$xo" ]; then if [ -z "$xonohome" -a ! -f $USBMNT/LiveOS/$HOMEFILE ]; then args="$args persistenthome=mtd0" fi + args="$args reset_overlay" if [ ! -d $USBMNT/boot ]; then mkdir -p $USBMNT/boot ; fi cat > $USBMNT/boot/olpc.fth < Date: Tue Oct 21 11:12:57 2008 -0400 Verify integer arguments to livecd-iso-to-disk (#467257) Verify that arguments passed to livecd-iso-to-disk are actually integers as opposed to specifying, eg, 256MB diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index 0a1144d..c224230 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -219,6 +219,12 @@ checkMounted() { fi } +checkint() { + if ! test $1 -gt 0 2>/dev/null ; then + usage + fi +} + if [ $(id -u) != 0 ]; then echo "You need to be root to run this script" exit 1 @@ -234,14 +240,17 @@ HOMEFILE="home.img" while [ $# -gt 2 ]; do case $1 in --overlay-size-mb) + checkint $2 overlaysizemb=$2 shift ;; --home-size-mb) + checkint $2 homesizemb=$2 shift ;; --swap-size-mb) + checkint $2 swapsizemb=$2 shift ;; From linux.charged at gmail.com Tue Oct 28 01:08:22 2008 From: linux.charged at gmail.com (Junior Tomazelli) Date: Mon, 27 Oct 2008 23:08:22 -0200 Subject: [Fedora-livecd-list] Copy files to livecd Message-ID: People, that directories are created on livecd, but I cant get copy some files of my disk to him. That is the command I used at the end of my kickstart: %post --nochroot Erro cp $INSTALL_ROOT/media/JuniorHD/Multim?dia/Imagens/Papeis de Parede/SolarVO.png $LIVE_ROOT/usr/share/wallpapers cp $INSTAL_ROOT/media/JuniorHD/Modificando/Kuai/Tema/Kuai.colors $LIVE_ROOT/usr/share/kde4/apps/color-schemes cp $INSTAL_ROOT/media/JuniorHD/Modificando/Kuai/Tema/GreenApparatus/* $LIVE_ROOT/home/Kuai/.icons/GreenApparatus %end -- Wolnei C?ndido Tomazelli Junior (Charged) Brazil Fedora Ambassador Designer and TI E-mail : linux.charged at gmail.com Linux User #477062 -------------- next part -------------- An HTML attachment was scrubbed... URL: From maillist at diffingo.com Tue Oct 28 03:03:56 2008 From: maillist at diffingo.com (Stewart Adam) Date: Mon, 27 Oct 2008 23:03:56 -0400 Subject: [Fedora-livecd-list] syslinux vs GRUB (Was: Using both EFI and MBR at once on Live USB images) In-Reply-To: <48F7FC16.9090201@diffingo.com> References: <48F7FC16.9090201@diffingo.com> Message-ID: <4906811C.7020803@diffingo.com> Hi, I've been testing out creating using a hybrid GPT+MBR stick so that EFI and BIOS-based systems can boot off of the one stick and share a persistent home. The good news is that it works and the changes are fairly simple - essentially, I just removed the conditional statements that prevented the non-mactel stuff from being executed when --mactell was passed. I also had to rearrange a few blocks so that the GPT partitions happens first, since the MBR should only be installed after GPT has initialized (and therefore it will install to GPT's protective MBR). The bad news is syslinux doesn't support GPT at the moment because we can't set a MBR partition as "bootable" in a GPT partition table. GRUB doesn't seem to care about this and will boot fine once installed to /dev/sdb, although the only bug I hit then was that booting from "USB-HDD" and "USB-ZIP" in the BIOS didn't work... I had to select "HDD" and then choose "USB-0". Is this why the scripts currently use syslinux instead of making the jump to GRUB? Even if we only use GRUB as a last resort (when --mactel is passed), I still think this is a worthfile feature most recent motherboards support booting from USB as a HDD instead of USB-ZIP. Stewart From bkearney at redhat.com Tue Oct 28 12:30:51 2008 From: bkearney at redhat.com (Bryan Kearney) Date: Tue, 28 Oct 2008 08:30:51 -0400 Subject: [Fedora-livecd-list] Copy files to livecd In-Reply-To: References: Message-ID: <490705FB.3030205@redhat.com> Is $LIVE_ROOT set outside of the chrooted environment? I am using a post section in some appliances that looks like this: %post --nochroot cp tcsplash.xpm.gz $INSTALL_ROOT/boot/grub/splash.xpm.gz %end Copying into $INSTALL_ROOT -- bk Junior Tomazelli wrote: > People, that directories are created on livecd, but I cant get copy some > files of my disk to him. That is the command I used at the end of my > kickstart: > > > %post --nochroot Erro > > cp $INSTALL_ROOT/media/JuniorHD/Multim?dia/Imagens/Papeis de > Parede/SolarVO.png $LIVE_ROOT/usr/share/wallpapers > cp $INSTAL_ROOT/media/JuniorHD/Modificando/Kuai/Tema/Kuai.colors > $LIVE_ROOT/usr/share/kde4/apps/color-schemes > cp $INSTAL_ROOT/media/JuniorHD/Modificando/Kuai/Tema/GreenApparatus/* > $LIVE_ROOT/home/Kuai/.icons/GreenApparatus > > %end > > -- > Wolnei C?ndido Tomazelli Junior (Charged) > Brazil Fedora Ambassador > Designer and TI > E-mail : linux.charged at gmail.com > Linux User #477062 > > > ------------------------------------------------------------------------ > > -- > Fedora-livecd-list mailing list > Fedora-livecd-list at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-livecd-list From katzj at redhat.com Tue Oct 28 15:11:10 2008 From: katzj at redhat.com (Jeremy Katz) Date: Tue, 28 Oct 2008 11:11:10 -0400 Subject: [Fedora-livecd-list] syslinux vs GRUB (Was: Using both EFI and MBR at once on Live USB images) In-Reply-To: <4906811C.7020803@diffingo.com> References: <48F7FC16.9090201@diffingo.com> <4906811C.7020803@diffingo.com> Message-ID: <1225206670.13184.6.camel@aglarond.local> On Mon, 2008-10-27 at 23:03 -0400, Stewart Adam wrote: > I've been testing out creating using a hybrid GPT+MBR stick so that EFI and > BIOS-based systems can boot off of the one stick and share a persistent > home. The good news is that it works and the changes are fairly simple - > essentially, I just removed the conditional statements that prevented the > non-mactel stuff from being executed when --mactell was passed. I also had > to rearrange a few blocks so that the GPT partitions happens first, since > the MBR should only be installed after GPT has initialized (and therefore it > will install to GPT's protective MBR). Cool! > The bad news is syslinux doesn't support GPT at the moment because we can't > set a MBR partition as "bootable" in a GPT partition table. GRUB doesn't > seem to care about this and will boot fine once installed to /dev/sdb, > although the only bug I hit then was that booting from "USB-HDD" and > "USB-ZIP" in the BIOS didn't work... I had to select "HDD" and then choose > "USB-0". Is this why the scripts currently use syslinux instead of making > the jump to GRUB? The biggest reason is that we use isolinux for CDs (because grub's cd booting code is incompatible with a fair number of machines) and using the same bootloader for both live image over CD and USB is a good thing. That said, syslinux definitely has better compatibility for booting off of various USB things since hpa has spent a lot of time dealing with BIOS bugs ;) It might be worth sending a mail to the syslinux list and see if hpa has any thoughts... he tends to be up for making syslinux as usable as possible for everywhere. > Even if we only use GRUB as a last resort (when --mactel is passed), I still > think this is a worthfile feature most recent motherboards support booting > from USB as a HDD instead of USB-ZIP. Yeah, if we have to, I guess we could fall back to this. But probably at this point, it's too late for Fedora 10 Jeremy From lmacken at fedoraproject.org Wed Oct 29 01:31:32 2008 From: lmacken at fedoraproject.org (Luke Macken) Date: Wed, 29 Oct 2008 01:31:32 +0000 (UTC) Subject: [Fedora-livecd-list] tools/livecd-iso-to-disk.sh Message-ID: <20081029013132.3CFA9120263@lists.fedorahosted.org> tools/livecd-iso-to-disk.sh | 3 +++ 1 file changed, 3 insertions(+) New commits: commit a0b104ebc118109a4768e3e2daf6d0c68c1def81 Author: Luke Macken Date: Tue Oct 28 21:30:17 2008 -0400 Add a --compress flag to allow compressed images along with --xo. diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh index 497062f..239846e 100755 --- a/tools/livecd-iso-to-disk.sh +++ b/tools/livecd-iso-to-disk.sh @@ -280,6 +280,9 @@ while [ $# -gt 2 ]; do --xo-no-home) xonohome=1 ;; + --compress) + skipcompress="" + ;; --skipcompress) skipcompress=1 ;; From apevec at redhat.com Wed Oct 29 18:57:43 2008 From: apevec at redhat.com (Alan Pevec) Date: Wed, 29 Oct 2008 19:57:43 +0100 Subject: [Fedora-livecd-list] Proposed patches for livecd/EL-5 branch Message-ID: <4908B227.3030300@redhat.com> Hi all, I have collected few backport patches for EPEL5 livecd, they're at http://repo.or.cz/w/livecd/EPEL5.git from tag livecd-tools-013-8.1-EPEL5-patches onwards. Please have a look, they're low risk and work for me. I'd like to push them to EPEL CVS if no objections. NB: the last one "add --global-config to use YUM global config" is a new feature which I'll post here for review shortly: http://repo.or.cz/w/livecd/EPEL5.git?a=commitdiff;h=65cea964999ac10d5f6660bcf7e7ff720994c91b;hp=9a7b6bd5faf9d3f006ccd60f8cc4d847dd03acc2 From sundaram at fedoraproject.org Wed Oct 29 19:17:38 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Thu, 30 Oct 2008 00:47:38 +0530 Subject: [Fedora-livecd-list] Re: Proposed patches for livecd/EL-5 branch In-Reply-To: <4908B227.3030300@redhat.com> References: <4908B227.3030300@redhat.com> Message-ID: <4908B6D2.10300@fedoraproject.org> Alan Pevec wrote: > Hi all, > > I have collected few backport patches for EPEL5 livecd, they're at > http://repo.or.cz/w/livecd/EPEL5.git from tag > livecd-tools-013-8.1-EPEL5-patches onwards. > > Please have a look, they're low risk and work for me. I'd like to push > them to EPEL CVS if no objections. Are these all backported fixes or are you adding more incompatible changes like the one below? I would like new features to be reviewed and accepted in the HEAD first and then backported to what lives in EPEL. Rahul > > NB: the last one "add --global-config to use YUM global config" is a new > feature which I'll post here for review shortly: > http://repo.or.cz/w/livecd/EPEL5.git?a=commitdiff;h=65cea964999ac10d5f6660bcf7e7ff720994c91b;hp=9a7b6bd5faf9d3f006ccd60f8cc4d847dd03acc2 > From apevec at redhat.com Wed Oct 29 20:29:37 2008 From: apevec at redhat.com (Alan Pevec) Date: Wed, 29 Oct 2008 21:29:37 +0100 Subject: [Fedora-livecd-list] Re: Proposed patches for livecd/EL-5 branch In-Reply-To: <4908B6D2.10300@fedoraproject.org> References: <4908B227.3030300@redhat.com> <4908B6D2.10300@fedoraproject.org> Message-ID: <4908C7B1.5040709@redhat.com> Rahul Sundaram wrote: > Alan Pevec wrote: >> Hi all, >> >> I have collected few backport patches for EPEL5 livecd, they're at >> http://repo.or.cz/w/livecd/EPEL5.git from tag >> livecd-tools-013-8.1-EPEL5-patches onwards. >> >> Please have a look, they're low risk and work for me. I'd like to push >> them to EPEL CVS if no objections. > > Are these all backported fixes or are you adding more incompatible all but last one are backports > changes like the one below? I would like new features to be reviewed and > accepted in the HEAD first and then backported to what lives in EPEL. yes, sorry if I wasn't clear, only backports go to EPEL CVS From sundaram at fedoraproject.org Wed Oct 29 20:46:43 2008 From: sundaram at fedoraproject.org (Rahul Sundaram) Date: Thu, 30 Oct 2008 02:16:43 +0530 Subject: [Fedora-livecd-list] Re: Proposed patches for livecd/EL-5 branch In-Reply-To: <4908C7B1.5040709@redhat.com> References: <4908B227.3030300@redhat.com> <4908B6D2.10300@fedoraproject.org> <4908C7B1.5040709@redhat.com> Message-ID: <4908CBB3.1040508@fedoraproject.org> Alan Pevec wrote: > yes, sorry if I wasn't clear, only backports go to EPEL CVS OK. Great. I am working remotely for a while without access to a EL system and I am running rawhide on my laptop for quite sometime. The connection is decent but not enough to start downloading new complete images easily so can't really test this at the moment but you can go ahead and commit those changes if you (or others) have tested that the following works correctly: * Create a live cd image * Use livecd-iso-to-disk script to convert the image into a bootable live USB * image-creator creates a ext3 image All fixes should be in HEAD first because incompatible changes are a headache when you add dependencies on them and you get to pick whether you want to keep them or move to the EL 6 branch when one gets created in the future. Thanks for working on this. Rahul From apevec at redhat.com Wed Oct 29 21:49:57 2008 From: apevec at redhat.com (Alan Pevec) Date: Wed, 29 Oct 2008 22:49:57 +0100 Subject: [Fedora-livecd-list] [PATCH] add --global-config Message-ID: <1225316997-11482-1-git-send-email-apevec@redhat.com> to use default YUM config i.e. /etc/yum.conf With this option you don't have to supply repo kickstart commands. Also add --cacheonly option to run entirely from YUM cache. Signed-off-by: Alan Pevec --- imgcreate/creator.py | 27 ++++++++++++++++++++------- imgcreate/yuminst.py | 11 +++++++---- tools/livecd-creator | 8 +++++++- 3 files changed, 34 insertions(+), 12 deletions(-) diff --git a/imgcreate/creator.py b/imgcreate/creator.py index 519735e..48d5d70 100644 --- a/imgcreate/creator.py +++ b/imgcreate/creator.py @@ -50,7 +50,7 @@ class ImageCreator(object): """ - def __init__(self, ks, name): + def __init__(self, ks, name, globalYUM = False, cacheonly = False): """Initialize an ImageCreator instance. ks -- a pykickstart.KickstartParser instance; this instance will be @@ -73,6 +73,9 @@ class ImageCreator(object): self.__builddir = None self.__bindmounts = [] + self.__globalYUM = globalYUM + self.__cacheonly = cacheonly + self.__sanity_check() def __del__(self): @@ -396,7 +399,8 @@ class ImageCreator(object): kickstart.convert_method_to_repo(self.ks) - if not kickstart.get_repos(self.ks): + # with -g global YUM config repositories are not taken from ks + if not self.__globalYUM and not kickstart.get_repos(self.ks): raise CreatorError("No repositories specified") if (kickstart.selinux_enabled(self.ks) and @@ -623,10 +627,17 @@ class ImageCreator(object): the kickstart to be overridden. """ - yum_conf = self._mktemp(prefix = "yum.conf-") ayum = LiveCDYum() - ayum.setup(yum_conf, self._instroot) + if self.__globalYUM: + yum_conf = None + else: + yum_conf = self._mktemp(prefix = "yum.conf-") + + if self.__cacheonly: + ayum.setup(yum_conf, self._instroot, cache=1) + else: + ayum.setup(yum_conf, self._instroot) for repo in kickstart.get_repos(self.ks, repo_urls): (name, baseurl, mirrorlist, inc, exc) = repo @@ -659,7 +670,8 @@ class ImageCreator(object): finally: ayum.closeRpmDB() ayum.close() - os.unlink(yum_conf) + if yum_conf: + os.unlink(yum_conf) # do some clean up to avoid lvm info leakage. this sucks. for subdir in ("cache", "backup", "archive"): @@ -776,7 +788,8 @@ class LoopImageCreator(ImageCreator): """ - def __init__(self, ks, name, fslabel = None): + def __init__(self, ks, name, fslabel = None, + globalYUM = False, cacheonly = False): """Initialize a LoopImageCreator instance. This method takes the same arguments as ImageCreator.__init__() with @@ -785,7 +798,7 @@ class LoopImageCreator(ImageCreator): fslabel -- A string used as a label for any filesystems created. """ - ImageCreator.__init__(self, ks, name) + ImageCreator.__init__(self, ks, name, globalYUM, cacheonly) self.__fslabel = None self.fslabel = fslabel diff --git a/imgcreate/yuminst.py b/imgcreate/yuminst.py index 04c9b40..c13b162 100644 --- a/imgcreate/yuminst.py +++ b/imgcreate/yuminst.py @@ -75,11 +75,14 @@ class LiveCDYum(yum.YumBase): for f in glob.glob(installroot + "/var/lib/rpm/__db*"): os.unlink(f) - def setup(self, confpath, installroot): - self._writeConf(confpath, installroot) + def setup(self, confpath, installroot, cache=0): self._cleanupRpmdbLocks(installroot) - self.doConfigSetup(fn = confpath, root = installroot) - self.conf.cache = 0 + if confpath: + self._writeConf(confpath, installroot) + self.doConfigSetup(fn = confpath, root = installroot) + else: + self.doConfigSetup(root = installroot) + self.conf.cache = cache self.doTsSetup() self.doRpmDBSetup() self.doRepoSetup() diff --git a/tools/livecd-creator b/tools/livecd-creator index b401190..594686a 100755 --- a/tools/livecd-creator +++ b/tools/livecd-creator @@ -52,6 +52,11 @@ def parse_options(args): sysopt.add_option("", "--cache", type="string", dest="cachedir", default=None, help="Cache directory to use (default: private cache") + sysopt.add_option("--global-config", dest="global_config", + action="store_true", + help="Use global YUM configuration, /etc/yum.conf") + sysopt.add_option("--cacheonly", action="store_true", dest="cacheonly", + help="run entirely from YUM cache, don't update YUM cache") parser.add_option_group(sysopt) imgcreate.setup_logging(parser) @@ -109,7 +114,8 @@ def main(): ks = imgcreate.read_kickstart(options.kscfg) - creator = imgcreate.LiveImageCreator(ks, name, fs_label) + creator = imgcreate.LiveImageCreator(ks, name, fs_label, + options.global_config, options.cacheonly) creator.tmpdir = options.tmpdir creator.skip_compression = options.skip_compression creator.skip_minimize = options.skip_minimize -- 1.5.6.5 From maillist at diffingo.com Wed Oct 29 22:18:12 2008 From: maillist at diffingo.com (Stewart Adam) Date: Wed, 29 Oct 2008 18:18:12 -0400 Subject: [Fedora-livecd-list] [PATCH] Using both EFI and MBR at once on Live USB images and other misc. fixes In-Reply-To: <48F8A696.8020601@diffingo.com> References: <48F7FC16.9090201@diffingo.com> <1224213914.370.85.camel@aglarond.local> <48F87E65.6040104@diffingo.com> <1224251419.13356.46.camel@aglarond.local> <48F8A696.8020601@diffingo.com> Message-ID: <4908E124.10603@diffingo.com> Hi, Here's the patch that I spoke about in my last email: http://downloads.diffingo.com/livecd-iso-to-disk.sh-efi.patch What it changes: * Exit messages when syslinux's MBR files can't be found * Unmounts $USBDEV after creating GPT, as HAL attempts to automount * Allows for hybrid (GPT+MBR) USB keys * Adds --skipcopy to skip copying the live OS * Adds --format option to replace --mactel + --reset-mbr * --reset-mbr will now reset the MBR, GPT or not. * if --format is specified, a valid partition isn't required (only a valid device instead) Notes worth mentioning: * --format + --mactel = new GPT table, resets PMBR (GPT's Protective MBR) * --format = new MSDOS table, resets MBR * syslinux support isn't complete let, hopefully coming soon * GPT hybrid tables requires /usr/{lib,share}/syslinux/gptmbr.bin which is only in the newest syslinux (not available in the Fedora repos yet). Hope this is useful! Stewart On 10/17/08 10:52 AM, Stewart Adam wrote: > Hm, I tried again and I got my errors mixed up. It actually says that > there's no data partition (only the EFI system one) so there's no need > to sync as the EFI system partition is always marked EFI Protective in > the MBR. If it isn't, then installing anything to the MBR risks erasing > the entire GPT data structure... > > So the after all the only option seems to be to use multiple partitions, > which I don't think will work well. I'm still going to try a few things > though... If GRUB can install to the MBR on the "EFI Protective" > partition and it realizes that "EFI Protective" is just FAT32, things > might work out on a single partition. > > Stewart > > -- > Fedora-livecd-list mailing list > Fedora-livecd-list at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-livecd-list > From maillist at diffingo.com Wed Oct 29 23:49:08 2008 From: maillist at diffingo.com (Stewart Adam) Date: Wed, 29 Oct 2008 19:49:08 -0400 Subject: [Fedora-livecd-list] [PATCH] Using both EFI and MBR at once on Live USB images and other misc. fixes In-Reply-To: <4908E124.10603@diffingo.com> References: <48F7FC16.9090201@diffingo.com> <1224213914.370.85.camel@aglarond.local> <48F87E65.6040104@diffingo.com> <1224251419.13356.46.camel@aglarond.local> <48F8A696.8020601@diffingo.com> <4908E124.10603@diffingo.com> Message-ID: <4908F674.6060408@diffingo.com> On 10/29/08 6:18 PM, Stewart Adam wrote: > * Adds --skipcopy to skip copying the live OS Oops, forgot to mention two things. I know this feature seems a little odd, so I wanted to explain a bit more... --skipcopy doesn't skip all copying, just the live OS which makes it useful for testing the boot configuration. It's also useful for USB media, where testing over and over would normally be wasting limited write cycles it takes a lot of waiting for the 700MB file copy finish. ALso, say the MBR gets lost or you've transferred the live OS files from a friend's key, you can just refresh the MBR and bootloader config instead without touching the live OS and persistant home. > * syslinux support isn't complete let, hopefully coming soon For this I mean GPT+MBR syslinux support, that made it sound like the patch broke the syslinux boot ;) Stewart From katzj at redhat.com Thu Oct 30 00:29:42 2008 From: katzj at redhat.com (Jeremy Katz) Date: Wed, 29 Oct 2008 20:29:42 -0400 Subject: [Fedora-livecd-list] [PATCH] add --global-config In-Reply-To: <1225316997-11482-1-git-send-email-apevec@redhat.com> References: <1225316997-11482-1-git-send-email-apevec@redhat.com> Message-ID: <1225326582.25818.1.camel@aglarond.local> On Wed, 2008-10-29 at 22:49 +0100, Alan Pevec wrote: > to use default YUM config i.e. /etc/yum.conf > With this option you don't have to supply repo kickstart commands. This makes it so that your images aren't reproducible without knowing what was installed on the system which was building the images. This circumvents one of the single biggest things about livecd-tools -- reproducibility > Also add --cacheonly option to run entirely from YUM cache. This could be okay, although what's the use case? Also, one patch per suggested change please. Jeremy From katzj at redhat.com Thu Oct 30 00:38:39 2008 From: katzj at redhat.com (Jeremy Katz) Date: Wed, 29 Oct 2008 20:38:39 -0400 Subject: [Fedora-livecd-list] [PATCH] Using both EFI and MBR at once on Live USB images and other misc. fixes In-Reply-To: <4908E124.10603@diffingo.com> References: <48F7FC16.9090201@diffingo.com> <1224213914.370.85.camel@aglarond.local> <48F87E65.6040104@diffingo.com> <1224251419.13356.46.camel@aglarond.local> <48F8A696.8020601@diffingo.com> <4908E124.10603@diffingo.com> Message-ID: <1225327119.25818.2.camel@aglarond.local> On Wed, 2008-10-29 at 18:18 -0400, Stewart Adam wrote: > Here's the patch that I spoke about in my last email: > http://downloads.diffingo.com/livecd-iso-to-disk.sh-efi.patch Cool. From a quick look, it looks okay. But given where we are in the Fedora 10 cycle, I'm going to hold off on applying it until after we get Fedora 10 out the door to avoid destabilizing things for people (especially as feature freeze was a month or so ago) Jeremy From maillist at diffingo.com Thu Oct 30 01:31:38 2008 From: maillist at diffingo.com (Stewart Adam) Date: Wed, 29 Oct 2008 21:31:38 -0400 Subject: [Fedora-livecd-list] [PATCH] Using both EFI and MBR at once on Live USB images and other misc. fixes In-Reply-To: <1225327119.25818.2.camel@aglarond.local> References: <48F7FC16.9090201@diffingo.com> <1224213914.370.85.camel@aglarond.local> <48F87E65.6040104@diffingo.com> <1224251419.13356.46.camel@aglarond.local> <48F8A696.8020601@diffingo.com> <4908E124.10603@diffingo.com> <1225327119.25818.2.camel@aglarond.local> Message-ID: <49090E7A.7090009@diffingo.com> On 10/29/08 8:38 PM, Jeremy Katz wrote: > On Wed, 2008-10-29 at 18:18 -0400, Stewart Adam wrote: >> Here's the patch that I spoke about in my last email: >> http://downloads.diffingo.com/livecd-iso-to-disk.sh-efi.patch > > Cool. From a quick look, it looks okay. But given where we are in the > Fedora 10 cycle, I'm going to hold off on applying it until after we get > Fedora 10 out the door to avoid destabilizing things for people > (especially as feature freeze was a month or so ago) > > Jeremy No problem, that gives us some more time to work out the GPT+syslinux problems. Stewart From apevec at redhat.com Thu Oct 30 11:35:16 2008 From: apevec at redhat.com (Alan Pevec) Date: Thu, 30 Oct 2008 12:35:16 +0100 Subject: [Fedora-livecd-list] [PATCH] add --global-config In-Reply-To: <1225326582.25818.1.camel@aglarond.local> References: <1225316997-11482-1-git-send-email-apevec@redhat.com> <1225326582.25818.1.camel@aglarond.local> Message-ID: <49099BF4.60502@redhat.com> Jeremy Katz wrote: > On Wed, 2008-10-29 at 22:49 +0100, Alan Pevec wrote: >> to use default YUM config i.e. /etc/yum.conf >> With this option you don't have to supply repo kickstart commands. > > This makes it so that your images aren't reproducible without knowing > what was installed on the system which was building the images. This > circumvents one of the single biggest things about livecd-tools -- > reproducibility Just knowing repo URLs doesn't help reproducibility - repo content can change. --global-config was added for experiments running livecd-creator in mock/koji/brew Mock adds repos from its config to /etc/yum.conf in the buildroot, and this, combined with exact koji repo ID, gets us real reproducibility. >> Also add --cacheonly option to run entirely from YUM cache. > > This could be okay, although what's the use case? Also, one patch per > suggested change please. Right, I'll do that. It should work like yum -C, the target use-case is to re-use mock yum_cache, where all packages from kickstart are pulled in by mock as BRs. See work-in-progress spec files, min-builder prepares buildroot and min is RPM wrapper for livecd ISO. -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: min-builder.spec URL: -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: min.spec URL: -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: min.ks URL: From katzj at redhat.com Thu Oct 30 14:50:21 2008 From: katzj at redhat.com (Jeremy Katz) Date: Thu, 30 Oct 2008 10:50:21 -0400 Subject: [Fedora-livecd-list] [PATCH] add --global-config In-Reply-To: <49099BF4.60502@redhat.com> References: <1225316997-11482-1-git-send-email-apevec@redhat.com> <1225326582.25818.1.camel@aglarond.local> <49099BF4.60502@redhat.com> Message-ID: <1225378221.25818.302.camel@aglarond.local> On Thu, 2008-10-30 at 12:35 +0100, Alan Pevec wrote: > Jeremy Katz wrote: > > On Wed, 2008-10-29 at 22:49 +0100, Alan Pevec wrote: > >> to use default YUM config i.e. /etc/yum.conf > >> With this option you don't have to supply repo kickstart commands. > > > > This makes it so that your images aren't reproducible without knowing > > what was installed on the system which was building the images. This > > circumvents one of the single biggest things about livecd-tools -- > > reproducibility > > Just knowing repo URLs doesn't help reproducibility - repo content can change. > --global-config was added for experiments running livecd-creator in mock/koji/brew > Mock adds repos from its config to /etc/yum.conf in the buildroot, and this, combined with exact koji repo ID, gets us real reproducibility. It helps quite a bit because you actually know what repos are being used as opposed to random stuff littered on the filesystem. Using the system yum config for livecd-creator is a non-starter. > >> Also add --cacheonly option to run entirely from YUM cache. > > > > This could be okay, although what's the use case? Also, one patch per > > suggested change please. > > Right, I'll do that. It should work like yum -C, the target use-case is to re-use mock yum_cache, > where all packages from kickstart are pulled in by mock as BRs. Umm... this really feels like going the wrong way to try to build livecds in mock for one thing. And reusing the cache can already work, just use --cachedir. Yes, if there are more things needed they'll get grabbed. But otherwise, it's just going to be a failure mode which I have a hard time seeing as "better" > See work-in-progress spec files, Umm, sudo for the builder in spec files? No thanks. If you're going to do that, you might as well just not use mock and build as root. Jeremy From bkearney at redhat.com Thu Oct 30 15:56:54 2008 From: bkearney at redhat.com (Bryan Kearney) Date: Thu, 30 Oct 2008 11:56:54 -0400 Subject: [Fedora-livecd-list] Re: Proposed patches for livecd/EL-5 branch In-Reply-To: <4908CBB3.1040508@fedoraproject.org> References: <4908B227.3030300@redhat.com> <4908B6D2.10300@fedoraproject.org> <4908C7B1.5040709@redhat.com> <4908CBB3.1040508@fedoraproject.org> Message-ID: <4909D946.2090102@redhat.com> Rahul Sundaram wrote: > Alan Pevec wrote: >> yes, sorry if I wasn't clear, only backports go to EPEL CVS > > OK. Great. I am working remotely for a while without access to a EL > system and I am running rawhide on my laptop for quite sometime. The > connection is decent but not enough to start downloading new complete > images easily so can't really test this at the moment but you can go > ahead and commit those changes if you (or others) have tested that the > following works correctly: > > * Create a live cd image > * Use livecd-iso-to-disk script to convert the image into a bootable > live USB > * image-creator creates a ext3 image I executed the first 2 using the attached kickstart file. I got a few warnings at boot up, but the images do work. Note.. the second one I did running the livecd-iso-to-disk from an F9 machine. Rahul.. what do you normally do for #3? -- bk -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: rhel5-live.ks URL: From orion at cora.nwra.com Thu Oct 30 23:08:26 2008 From: orion at cora.nwra.com (Orion Poplawski) Date: Thu, 30 Oct 2008 17:08:26 -0600 Subject: [Fedora-livecd-list] Trouble mounting persistent home... In-Reply-To: <49010991.5040406@cora.nwra.com> References: <49010733.3080301@cora.nwra.com> <49010793.3090508@cora.nwra.com> <49010991.5040406@cora.nwra.com> Message-ID: <490A3E6A.3010708@cora.nwra.com> Orion Poplawski wrote: > > Okay, fedora-live does: > > mount -o ro /dev/live /mnt/live > > if I do: > > mount /dev/live /mnt/live > > it works. Forget the fuser stuff, that's me being dumb. > > So, what's causing this? > Filed https://bugzilla.redhat.com/show_bug.cgi?id=469269 -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion at cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com From bkearney at redhat.com Fri Oct 31 13:51:38 2008 From: bkearney at redhat.com (Bryan Kearney) Date: Fri, 31 Oct 2008 09:51:38 -0400 Subject: [Fedora-livecd-list] Re: Proposed patches for livecd/EL-5 branch In-Reply-To: <4909D946.2090102@redhat.com> References: <4908B227.3030300@redhat.com> <4908B6D2.10300@fedoraproject.org> <4908C7B1.5040709@redhat.com> <4908CBB3.1040508@fedoraproject.org> <4909D946.2090102@redhat.com> Message-ID: <490B0D6A.3010604@redhat.com> Bryan Kearney wrote: > Rahul Sundaram wrote: >> Alan Pevec wrote: >>> yes, sorry if I wasn't clear, only backports go to EPEL CVS >> >> OK. Great. I am working remotely for a while without access to a EL >> system and I am running rawhide on my laptop for quite sometime. The >> connection is decent but not enough to start downloading new complete >> images easily so can't really test this at the moment but you can go >> ahead and commit those changes if you (or others) have tested that the >> following works correctly: >> >> * Create a live cd image >> * Use livecd-iso-to-disk script to convert the image into a bootable >> live USB >> * image-creator creates a ext3 image > > > I executed the first 2 using the attached kickstart file. I got a few > warnings at boot up, but the images do work. Note.. the second one I did > running the livecd-iso-to-disk from an F9 machine. > > Rahul.. what do you normally do for #3? > > -- bk > I verified the livecd-iso-to-disk on a rhel 5.2 machine. Worked fine for me. -- bk