From leilei175 at gmail.com Fri Apr 2 07:53:07 2010 From: leilei175 at gmail.com (leilei175 at gmail.com) Date: Fri, 2 Apr 2010 15:53:07 +0800 Subject: http/ftp install fail but nfs works Message-ID: Hi, guys I am trying to install RedHat EL4 remotely. The way I tried is insert the disc_1 to boot the raw box -> using dhcp to start network -> install method selection When I select http or ftp method and input the http/ftp website name and directory, it will pop up a dialog box asking me to test the CD media or skip. when I choose skip it would went back to the http setup stage(asking for website name and directory). This comes to an dead loop. The network connection should be fine. I exported the installlation directory on the same server ,and nfs install went well. Is there anything I'm doing wrong to use http/ftp install method? Thanks Lei From crhea at mayo.edu Fri Apr 2 16:46:57 2010 From: crhea at mayo.edu (Cris Rhea) Date: Fri, 2 Apr 2010 11:46:57 -0500 Subject: FTP/HTTP vs NFS install In-Reply-To: References: Message-ID: <20100402164657.GB18926@kaizen.mayo.edu> > Hi, guys > I am trying to install RedHat EL4 remotely. > The way I tried is > insert the disc_1 to boot the raw box -> using dhcp to start network > -> install method selection > > When I select http or ftp method and input the http/ftp website name > and directory, it will pop up a dialog box asking me to test the CD > media or skip. > when I choose skip it would went back to the http setup stage(asking > for website name and directory). > This comes to an dead loop. > > The network connection should be fine. I exported the installlation > directory on the same server ,and nfs install went well. > > Is there anything I'm doing wrong to use http/ftp install method? > > Thanks > Lei When you get to the initial screen, are you entering: linux askmethod There is some goofiness with kickstart if you have a CD loaded, it keeps assuming you want to use it. For this sort of situation, I also tend to use the netboot.iso image, rather than disc_1 image. These days, I use PXE boot (for cluster nodes) or boot the netboot.iso image via a USB memory key. For both of these, I use the FTP method. -- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From leilei175 at gmail.com Sun Apr 4 08:56:39 2010 From: leilei175 at gmail.com (leilei175 at gmail.com) Date: Sun, 4 Apr 2010 16:56:39 +0800 Subject: FTP/HTTP vs NFS install In-Reply-To: <20100402164657.GB18926@kaizen.mayo.edu> References: <20100402164657.GB18926@kaizen.mayo.edu> Message-ID: Thanks for your help, Cris Yes, I do use "linux askmethod" on the command line. So if I insert a CD to boot the raw box, it is impossible to use remote http/ftp install trees? However I was able to use nfs in this situation. That's weird. I don't have PXE in my environment If I would like to boot a raw box and ask it to use kickstart file and install tree on a http server what should I do, Could you give me some suggestions? I havn't used the netboot.iso image you mentioned. Is it possible to create a bootable CD with netboot.iso? Thanks Lei 2010/4/3 Cris Rhea : >> Hi, guys >> I am trying to install RedHat EL4 remotely. >> The way I tried is >> insert the disc_1 to boot the raw box -> using dhcp to start network >> -> install method selection >> >> When I select http or ftp method and input the http/ftp website name >> and directory, it will pop up a dialog box asking me to test the CD >> media or skip. >> when I choose skip it would went back to the http setup stage(asking >> for website name and directory). >> This comes to an dead loop. >> >> The network connection should be fine. I exported the installlation >> directory on the same server ,and nfs install went well. >> >> Is there anything I'm doing wrong to use http/ftp install method? >> >> Thanks >> Lei > > When you get to the initial screen, are you entering: > > linux askmethod > > There is some goofiness with kickstart if you have a CD loaded, it > keeps assuming you want to use it. For this sort of situation, > I also tend to use the netboot.iso image, rather than disc_1 image. > > These days, I use PXE boot (for cluster nodes) or boot the netboot.iso > image via a USB memory key. ?For both of these, I use the FTP method. > > -- Cris > > -- > ?Cristopher J. Rhea > ?Mayo Clinic - Research Computing Facility > ?200 First St SW, Rochester, MN 55905 > ?crhea at Mayo.EDU > ?(507) 284-0587 > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From sergio at sergiomb.no-ip.org Sun Apr 4 13:25:47 2010 From: sergio at sergiomb.no-ip.org (Sergio Monteiro Basto) Date: Sun, 04 Apr 2010 14:25:47 +0100 Subject: FTP/HTTP vs NFS install In-Reply-To: References: <20100402164657.GB18926@kaizen.mayo.edu> Message-ID: <1270387547.4290.5.camel@segulix> On Sun, 2010-04-04 at 16:56 +0800, leilei175 at gmail.com wrote: > Is it possible to create a bootable CD with netboot.iso? with netboot.iso I don't know, but with boot.iso you can do a bootable pen usb. if pen usb is in /dev/sdb1 umount /dev/sdb1 mkfs.vfat /dev/sdb1 /usr/bin/livecd-iso-to-disk boot.iso /dev/sdb1 # This I think is already done automatically in fedora 12. mount /dev/sdb1 /media/disk/ mkdir /media/disk/images/ cp install.img /media/disk/images/ -- S?rgio M. B. -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3159 bytes Desc: not available URL: From jason at rampaginggeek.com Sun Apr 4 14:56:27 2010 From: jason at rampaginggeek.com (Jason Edgecombe) Date: Sun, 04 Apr 2010 10:56:27 -0400 Subject: FTP/HTTP vs NFS install In-Reply-To: References: <20100402164657.GB18926@kaizen.mayo.edu> Message-ID: <4BB8A89B.2060302@rampaginggeek.com> leilei175 at gmail.com wrote: > Thanks for your help, Cris > > Yes, I do use "linux askmethod" on the command line. > So if I insert a CD to boot the raw box, it is impossible to use > remote http/ftp install trees? > However I was able to use nfs in this situation. That's weird. > > I don't have PXE in my environment > If I would like to boot a raw box and ask it to use kickstart file and > install tree on a http server > what should I do, Could you give me some suggestions? > > I havn't used the netboot.iso image you mentioned. > Is it possible to create a bootable CD with netboot.iso? > I'm not sure if this will work for RHEL4, but for RHEL5, I can just append my ks=http://www.example.com/kickstart.txt at the boot prompt. Similarly, I burned a custom build CD and a USB key: rough instructions for CD: 1. copy isolinux/vmlinux, vmlinux/initrd, and isolinux.isolinux.bin from the first RHEL5 CD to a /tmp/src. 2. copy isolinux/isolinux.cfg to /tmp/src and customize it by adding your custom ks line and removing uneeded stuff 3. create /tmp/src/boot.msg and put some text to describe what to type 4. run mkbootcd ============mkbootcd============== cd /tmp/src/ mkisofs -o ../boot.iso \ -b isolinux.bin -c boot.cat \ -no-emul-boot -boot-load-size 4 -boot-info-table \ ./ cd .. ================================ burn and test /tmp/boot.iso to make a pen drive, run "syslinux dev" on a fat32-formatted thumb drive ( substitute dev for the block device of the thumb drive), copy /tmp/src/* to the drive and rename isolinux.cfg to syslinux.cfg I use this method to build RHEL5 boxes which fetch the kickstart file over http and install from NFS Jason From crhea at mayo.edu Sun Apr 4 16:30:49 2010 From: crhea at mayo.edu (Cris Rhea) Date: Sun, 4 Apr 2010 11:30:49 -0500 Subject: FTP/HTTP vs NFS install In-Reply-To: References: Message-ID: <20100404163049.GA16905@kaizen.mayo.edu> > Thanks for your help, Cris > > Yes, I do use "linux askmethod" on the command line. > So if I insert a CD to boot the raw box, it is impossible to use > remote http/ftp install trees? If you are using the "linux askmethod", YES, you should be able to boot from CD and use http/ftp install trees. I've done it for years (just moved to USB memory sticks as they are more convenient). > I don't have PXE in my environment > If I would like to boot a raw box and ask it to use kickstart file and > install tree on a http server > what should I do, Could you give me some suggestions? I think you're on the right track. Is there any error reported in the KS logs or on the HTTP server? While I typically use FTP, there's virtually no difference in using HTTP. > I havn't used the netboot.iso image you mentioned. > Is it possible to create a bootable CD with netboot.iso? Yes, the netboot.iso is specifically made to make a bootable CD and do what you are trying to do. Just burn it to a CD, like any other iso. > > Thanks > Lei --- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From leilei175 at gmail.com Tue Apr 6 06:25:15 2010 From: leilei175 at gmail.com (leilei175 at gmail.com) Date: Tue, 6 Apr 2010 14:25:15 +0800 Subject: FTP/HTTP vs NFS install In-Reply-To: <20100404163049.GA16905@kaizen.mayo.edu> References: <20100404163049.GA16905@kaizen.mayo.edu> Message-ID: this seems to be a bug. My procedure is 1. insert the first CD to boot the box 2. use "linux askmethod " to choose the install method (cdrom/nfs/http/ftp) 3. choose http method 4. enter the http server hostname and directory name 5. it pop up the CD media check window asking me whether to do the media check or skip it 6. I choose skip and it would jump back to the http server/dir name screen then it went to 5 again and went into a dead loop. If I use initrd.img and vmlinz on the command line to boot, everything went well. I choose http method, entering the server hostname and directory name, it would proceed to installation. Has anyone encountered this problem before? Btw,I'm using a vmware guest to test this, dhcp is used to get network. Thanks 2010/4/5, Cris Rhea : >> Thanks for your help, Cris >> >> Yes, I do use "linux askmethod" on the command line. >> So if I insert a CD to boot the raw box, it is impossible to use >> remote http/ftp install trees? > > If you are using the "linux askmethod", YES, you should be able to > boot from CD and use http/ftp install trees. > > I've done it for years (just moved to USB memory sticks as they are > more convenient). > >> I don't have PXE in my environment >> If I would like to boot a raw box and ask it to use kickstart file and >> install tree on a http server >> what should I do, Could you give me some suggestions? > > I think you're on the right track. Is there any error reported > in the KS logs or on the HTTP server? > > While I typically use FTP, there's virtually no difference in using HTTP. > >> I havn't used the netboot.iso image you mentioned. >> Is it possible to create a bootable CD with netboot.iso? > > Yes, the netboot.iso is specifically made to make a bootable CD > and do what you are trying to do. Just burn it to a CD, like any other > iso. > > >> >> Thanks >> Lei > > > --- Cris > > -- > Cristopher J. Rhea > Mayo Clinic - Research Computing Facility > 200 First St SW, Rochester, MN 55905 > crhea at Mayo.EDU > (507) 284-0587 > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From dxh at yahoo.com Tue Apr 27 03:03:34 2010 From: dxh at yahoo.com (Don Hoover) Date: Mon, 26 Apr 2010 20:03:34 -0700 (PDT) Subject: Kickstart issues in RHEL6 Beta Message-ID: <243025.15141.qm@web65503.mail.ac4.yahoo.com> CC: to kickstart list for people like me who do lots of %pre/%post stuff. A couple of issues I discovered when I just threw EL6 Beta into my existing EL5 kickstart. 1) /tmp/netinfo is now gone. This is one of the most widely documented place to easily get the network info when doing a %pre or %post script. I guess I will write some new code to get the info manually out of ifconfig etc.. but a note about this would be nice in the EL6 release notes since lots of people probably use this today in EL5 kickstart scripts. 2) In the current EL6 beta, the "list-harddrives" command inside the anaconda environment is totally broken if you have a CDROM drive. It crashes with an error of "unknown disk label" from the python code. I believe that this little utility was widely used in %pre and %post scripts as well. It was the quickest way to get a list of all the disks on the system and their info. I used it to verify the correct drive sizes and to automatically choose the first correct disk to install to in my kickstart. I even used a formula based on the disk size info from list-harddrives to calculate the size of the swap lv. These sort of things were what many of us used that built hands-off fully automated installs. If they are going away in EL6, they should be documented I would think. From clumens at redhat.com Tue Apr 27 03:55:44 2010 From: clumens at redhat.com (Chris Lumens) Date: Mon, 26 Apr 2010 23:55:44 -0400 Subject: Kickstart issues in RHEL6 Beta In-Reply-To: <243025.15141.qm@web65503.mail.ac4.yahoo.com> References: <243025.15141.qm@web65503.mail.ac4.yahoo.com> Message-ID: <20100427035544.GO2546@localhost.localdomain> > 1) /tmp/netinfo is now gone. This is one of the most widely > documented place to easily get the network info when doing a %pre or > %post script. I guess I will write some new code to get the info > manually out of ifconfig etc.. but a note about this would be nice in > the EL6 release notes since lots of people probably use this today in > EL5 kickstart scripts. /tmp/netinfo was an implementation detail and a bit of a hack - really, a way for anaconda's stage2 to know what loader set up. loader doesn't set up networking anymore. NetworkManager does that. Therefore, we don't require netinfo anymore. anaconda can get all that from dbus instead of having to write out a file and read it back in again later. > 2) In the current EL6 beta, the "list-harddrives" command inside the > anaconda environment is totally broken if you have a CDROM drive. It > crashes with an error of "unknown disk label" from the python code. I > believe that this little utility was widely used in %pre and %post > scripts as well. It was the quickest way to get a list of all the > disks on the system and their info. I used it to verify the correct > drive sizes and to automatically choose the first correct disk to > install to in my kickstart. I even used a formula based on the disk > size info from list-harddrives to calculate the size of the swap lv. File a bug, but please don't use any of the command stubs unless you absolutely have to. They do not recieve much of any testing or fixing. - Chris From bschneiders at woti.com Tue Apr 27 14:13:18 2010 From: bschneiders at woti.com (Bryan Schneiders) Date: Tue, 27 Apr 2010 10:13:18 -0400 Subject: Kickstart issues in RHEL6 Beta In-Reply-To: <20100427035544.GO2546@localhost.localdomain> References: <243025.15141.qm@web65503.mail.ac4.yahoo.com> <20100427035544.GO2546@localhost.localdomain> Message-ID: <4BD6F0FE.9010603@woti.com> One of the snippets in cobbler moved away from list-harddrives to using /proc/partitions to be more reliable. Their old code looked like this: -drives=$(list-harddrives | awk '{print $1}') -for disk in $drives; do - DISKS="$DISKS $(fdisk -l /dev/$disk | awk '/^\/dev/{print $1}')" -done Their new code looks like this: +DISKS=$(awk '{if ($NF ~ "^[a-zA-Z].*[0-9]$" && $NF !~ "c[0-9]+d[0-9]+$") print "/dev/"$NF}' /proc/partitions) +# In the awk line above we want to make list of partitions, but not devices/controllers +# cciss raid controllers have partitions like /dev/cciss/cNdMpL, where N,M,L - some digits, we want to make sure 'pL' is there +# I hope this helps. The leading - and + are from the patch. Bryan Schneiders On 04/26/2010 11:55 PM, Chris Lumens wrote: >> 1) /tmp/netinfo is now gone. This is one of the most widely >> documented place to easily get the network info when doing a %pre or >> %post script. I guess I will write some new code to get the info >> manually out of ifconfig etc.. but a note about this would be nice in >> the EL6 release notes since lots of people probably use this today in >> EL5 kickstart scripts. > > /tmp/netinfo was an implementation detail and a bit of a hack - really, > a way for anaconda's stage2 to know what loader set up. loader doesn't > set up networking anymore. NetworkManager does that. Therefore, we > don't require netinfo anymore. anaconda can get all that from dbus > instead of having to write out a file and read it back in again later. > >> 2) In the current EL6 beta, the "list-harddrives" command inside the >> anaconda environment is totally broken if you have a CDROM drive. It >> crashes with an error of "unknown disk label" from the python code. I >> believe that this little utility was widely used in %pre and %post >> scripts as well. It was the quickest way to get a list of all the >> disks on the system and their info. I used it to verify the correct >> drive sizes and to automatically choose the first correct disk to >> install to in my kickstart. I even used a formula based on the disk >> size info from list-harddrives to calculate the size of the swap lv. > > File a bug, but please don't use any of the command stubs unless you > absolutely have to. They do not recieve much of any testing or fixing. > > - Chris > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From dxh at yahoo.com Tue Apr 27 18:36:53 2010 From: dxh at yahoo.com (Don Hoover) Date: Tue, 27 Apr 2010 11:36:53 -0700 (PDT) Subject: Kickstart issues in RHEL6 Beta Message-ID: <154762.43975.qm@web65515.mail.ac4.yahoo.com> Yes, I have written very similar code to parse /proc/partitions myself as a substitute. I opened bugzilla# 586410 about the error anyway. After all if the command is still in /usr/sbin it should work. I also opened bugzilla# 586188 about putting a note about the removal of /tmp/netinfo into the release notes for RHEL6. And I opened TWO NEW bugzilla's today, both of these are releated to busybox being removed and a couple of common functions are no longer available. 2) Bugzilla# 586499 - There is no longer a 'clear' screen command available in the kickstart environment for %pre/%post scripts This can be accomplished via echo -e "\033[H\033[J" 1) Bugzilla# 586494 - The ever popular chvt command is no longer exists for use in %pre/%post scripts. I have no idea how to do this yet without the chvt command. On the bright side... I ALMOST have gotten my RHEL5 kickstart to actually not crash yet. From mjenning at rackspace.com Tue Apr 27 19:21:00 2010 From: mjenning at rackspace.com (Michael Jennings) Date: Tue, 27 Apr 2010 14:21:00 -0500 Subject: Kickstart issues in RHEL6 Beta In-Reply-To: <154762.43975.qm@web65515.mail.ac4.yahoo.com> References: <154762.43975.qm@web65515.mail.ac4.yahoo.com> Message-ID: <8449_1272396067_o3RJL1jD008730_167D46F1-0C05-44F5-9B1D-FE616CAF6C75@rackspace.com> On Apr 27, 2010, at 1:36 PM, Don Hoover wrote: > Yes, I have written very similar code to parse /proc/partitions myself as a substitute. > > I opened bugzilla# 586410 about the error anyway. After all if the command is still in /usr/sbin it should work. > > I also opened bugzilla# 586188 about putting a note about the removal of /tmp/netinfo into the release notes for RHEL6. > > > And I opened TWO NEW bugzilla's today, both of these are releated to busybox being removed and a couple of common functions are no longer available. > > 2) Bugzilla# 586499 - There is no longer a 'clear' screen command available in the kickstart environment for %pre/%post scripts > > This can be accomplished via echo -e "\033[H\033[J" > > > > 1) Bugzilla# 586494 - The ever popular chvt command is no longer exists for use in %pre/%post scripts. > > I have no idea how to do this yet without the chvt command. You do have openvt I noticed. Try using that. I want chvt readded as well though. > > > > > On the bright side... I ALMOST have gotten my RHEL5 kickstart to actually not crash yet. > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list Confidentiality Notice: This e-mail message (including any attached or embedded documents) is intended for the exclusive and confidential use of the individual or entity to which this message is addressed, and unless otherwise expressly indicated, is confidential and privileged information of Rackspace. Any dissemination, distribution or copying of the enclosed material is prohibited. If you receive this transmission in error, please notify us immediately by e-mail at abuse at rackspace.com, and delete the original message. Your cooperation is appreciated. From lijian.gnu at gmail.com Wed Apr 28 10:43:29 2010 From: lijian.gnu at gmail.com (Jian Lee) Date: Wed, 28 Apr 2010 18:43:29 +0800 Subject: where is the "BaseHandler.__init__" exec ? Message-ID: <201004281843.29915.lijian.gnu@gmail.com> hi,all I'm research pykickstart now, have a puzzle on pykickstart/version.py . where is the "BaseHandler.__init__" exec ? look at the following code, after func "returnClassForVersion" have executed, the "commands" attr of "handler" have fixed by "command-object" mapping already. but from the line 173 to 177, there is just a imputil.imp.load_module func , and the module itself haven't execute a "BaseHandler.__init__" func. how the "handler" get a fixed attr "commands" ? Thanks, ------------------------------------------------------------------- 159 def returnClassForVersion(version=DEVEL): 160 """Return the class of the syntax handler for version. version can be 161 either a string or the matching constant. Raises KickstartValueError 162 if version does not match anything. 163 """ 164 try: 165 version = int(version) 166 module = "%s" % versionToString(version, skipDevel=True) 167 except ValueError: 168 module = "%s" % version 169 version = stringToVersion(version) 170 171 module = module.lower() 172 173 try: 174 import pykickstart.handlers 175 sys.path.extend(pykickstart.handlers.__path__) 176 found = imputil.imp.find_module(module) 177 loaded = imputil.imp.load_module(module, found[0], found[1], found[2]) 178 179 for (k, v) in loaded.__dict__.iteritems(): 180 if k.lower().endswith("%shandler" % module): 181 return v 182 except: 183 raise KickstartVersionError(_("Unsupported version specified: %s") % version) ===================================================== --------------------- Jian Lee [ http://jianlee.ylinux.org ] From clumens at redhat.com Wed Apr 28 14:04:49 2010 From: clumens at redhat.com (Chris Lumens) Date: Wed, 28 Apr 2010 10:04:49 -0400 Subject: where is the "BaseHandler.__init__" exec ? In-Reply-To: <201004281843.29915.lijian.gnu@gmail.com> References: <201004281843.29915.lijian.gnu@gmail.com> Message-ID: <20100428140449.GR2546@localhost.localdomain> > where is the "BaseHandler.__init__" exec ? > > look at the following code, after func "returnClassForVersion" have executed, > the "commands" attr of "handler" have fixed by "command-object" mapping already. > > but from the line 173 to 177, there is just a imputil.imp.load_module func , > and the module itself haven't execute a "BaseHandler.__init__" func. > > how the "handler" get a fixed attr "commands" ? returnClassForVersion just returns the class. It does not return an instance of that class, so you have to do that yourself. Look at makeVersion in pykickstart/version.py. That returns an instance of the class, and that's how the handler gets a commands dict. Look at pykickstart/base.py:BaseHandler.__init__. There, _registerCommands is called which sets up that attr. Does that clear things up? - Chris From kmcdonald at egenera.com Wed Apr 28 18:15:29 2010 From: kmcdonald at egenera.com (Kyle McDonald) Date: Wed, 28 Apr 2010 14:15:29 -0400 Subject: Can the NFS media location be passed to anaconda as a kernel arg? Message-ID: <4BD87B41.8070908@egenera.com> Hi All, I want to setup PXE network installs that are interactive, but I'd like to limit the amount of questions the user has to enter for each one. The PXE boot already passes some kernel args to configure the serial console, and force a vnc install. I'm wondering if I can also supply the NFS location for the media as a kernel arg so that users don't need to type it in everytime (It's always the same, and it's long.) Also I know when doing KickStart installs, you can use ks-device=eth1 to avoid being prompted for the network device when there is more than one. Can I use this same option so that the interactive user doesn't have to know which eth to pick? On a related note, is there a way to have anaconda figure out which interface the PXE boot booted the machine on and just go with that one? Right now I need to make 2 different PXE boot config files for the 2 types of machines I have because one boots on eth0 and the other on eth1. Thanks, -Kyle From wes.hardin at maxim-ic.com Wed Apr 28 20:36:05 2010 From: wes.hardin at maxim-ic.com (Wes Hardin) Date: Wed, 28 Apr 2010 15:36:05 -0500 Subject: Can the NFS media location be passed to anaconda as a kernel arg? In-Reply-To: <4BD87B41.8070908@egenera.com> References: <4BD87B41.8070908@egenera.com> Message-ID: <4BD89C35.3090103@maxim-ic.com> On 04/28/2010 01:15 PM, Kyle McDonald wrote: > I want to setup PXE network installs that are interactive, but I'd like > to limit the amount of questions the user has to enter for each one. The > PXE boot already passes some kernel args to configure the serial > console, and force a vnc install. I'm wondering if I can also supply the > NFS location for the media as a kernel arg so that users don't need to > type it in everytime (It's always the same, and it's long.) > > Also I know when doing KickStart installs, you can use ks-device=eth1 to > avoid being prompted for the network device when there is more than one. > Can I use this same option so that the interactive user doesn't have to > know which eth to pick? > > On a related note, is there a way to have anaconda figure out which > interface the PXE boot booted the machine on and just go with that one? > Right now I need to make 2 different PXE boot config files for the 2 > types of machines I have because one boots on eth0 and the other on eth1. I don't think media location is something you pass in on the kernel line. Instead, what you can pass in is the path to a kickstart file which then answers a lot of those repetitive questions, such as the location of the install tree. The kickstart file can be as complete or as minimal as you wish. I have one that only contains the network location of the install tree and waits for the installer to answer the rest. As for using the same interface that you PXE'd from, check out the ksdevice=bootif option, described briefly here: http://wiki.centos.org/TipsAndTricks/KickStart This document from Dell mentions it too: http://linux.dell.com/files/whitepapers/nic-enum-whitepaper-v2.pdf I don't claim to have figured this out myself, it's set by default on all my Cobbler PXE profiles. /* Wes Hardin */ From wes.hardin at maxim-ic.com Wed Apr 28 20:47:39 2010 From: wes.hardin at maxim-ic.com (Wes Hardin) Date: Wed, 28 Apr 2010 15:47:39 -0500 Subject: Can the NFS media location be passed to anaconda as a kernel arg? In-Reply-To: <4BD89C35.3090103@maxim-ic.com> References: <4BD87B41.8070908@egenera.com> <4BD89C35.3090103@maxim-ic.com> Message-ID: <4BD89EEB.1030705@maxim-ic.com> On 04/28/2010 03:36 PM, Wes Hardin wrote: > On 04/28/2010 01:15 PM, Kyle McDonald wrote: >> I want to setup PXE network installs that are interactive, but I'd like >> to limit the amount of questions the user has to enter for each one. The >> PXE boot already passes some kernel args to configure the serial >> console, and force a vnc install. I'm wondering if I can also supply the >> NFS location for the media as a kernel arg so that users don't need to >> type it in everytime (It's always the same, and it's long.) >> >> Also I know when doing KickStart installs, you can use ks-device=eth1 to >> avoid being prompted for the network device when there is more than one. >> Can I use this same option so that the interactive user doesn't have to >> know which eth to pick? >> >> On a related note, is there a way to have anaconda figure out which >> interface the PXE boot booted the machine on and just go with that one? >> Right now I need to make 2 different PXE boot config files for the 2 >> types of machines I have because one boots on eth0 and the other on eth1. > > I don't think media location is something you pass in on the kernel line. > Instead, what you can pass in is the path to a kickstart file which then answers > a lot of those repetitive questions, such as the location of the install tree. > The kickstart file can be as complete or as minimal as you wish. I have one > that only contains the network location of the install tree and waits for the > installer to answer the rest. > > As for using the same interface that you PXE'd from, check out the > ksdevice=bootif option, described briefly here: > > http://wiki.centos.org/TipsAndTricks/KickStart > > This document from Dell mentions it too: > > http://linux.dell.com/files/whitepapers/nic-enum-whitepaper-v2.pdf > > I don't claim to have figured this out myself, it's set by default on all my > Cobbler PXE profiles. And just as a note, I don't think this works on RHEL4. I had a workstation last week with 2 interfaces that I ended up just disabling the second interface to get it to work sanely. Here's the note on the change from the Cobbler folks, which includes a link to the relevant Fedora 9 documentation: https://fedorahosted.org/pipermail/cobbler/2008-September/000824.html -- /* Wes Hardin */ From kmcdonald at egenera.com Wed Apr 28 21:05:42 2010 From: kmcdonald at egenera.com (Kyle McDonald) Date: Wed, 28 Apr 2010 17:05:42 -0400 Subject: Can the NFS media location be passed to anaconda as a kernel arg? In-Reply-To: <4BD89C35.3090103@maxim-ic.com> References: <4BD87B41.8070908@egenera.com> <4BD89C35.3090103@maxim-ic.com> Message-ID: <4BD8A326.7090604@egenera.com> On 4/28/2010 4:36 PM, Wes Hardin wrote: > On 04/28/2010 01:15 PM, Kyle McDonald wrote: >> I want to setup PXE network installs that are interactive, but I'd like >> to limit the amount of questions the user has to enter for each one. The >> PXE boot already passes some kernel args to configure the serial >> console, and force a vnc install. I'm wondering if I can also supply the >> NFS location for the media as a kernel arg so that users don't need to >> type it in everytime (It's always the same, and it's long.) >> >> Also I know when doing KickStart installs, you can use ks-device=eth1 to >> avoid being prompted for the network device when there is more than one. >> Can I use this same option so that the interactive user doesn't have to >> know which eth to pick? >> >> On a related note, is there a way to have anaconda figure out which >> interface the PXE boot booted the machine on and just go with that one? >> Right now I need to make 2 different PXE boot config files for the 2 >> types of machines I have because one boots on eth0 and the other on >> eth1. > > I don't think media location is something you pass in on the kernel > line. Instead, what you can pass in is the path to a kickstart file > which then answers a lot of those repetitive questions, such as the > location of the install tree. The kickstart file can be as complete or > as minimal as you wish. I have one that only contains the network > location of the install tree and waits for the installer to answer the > rest. That blows. I already have a PXE config file for each release of RHEL SLES and Solaris I need to install. These are required to direct the network boot to the right release. For SLES and Solaris, I can put the media location in this PXE config file along with the install config file (AutoYast or JumpStart), and I can get away with just 1 install config file for all my machines (I use the begin and finish scripts to tweak the profile based on what release was booted so 1 file really helps me out since when the I do need to make a code update I only need to edit it once.) But for RHEL (and CENTOS) I have to have 2 (32 and 64 bit) KS config files for each release that only differ by the 'nfs ....' line. Now If I want to have an interactive install where the media location is already set, I need to go and make 2 more KS files for every release? That's # releases * 4 files to update when and if the rest of the file needs to change. And the 'nfs ...' line is one of the things that the %pre can't generate when it writes out the file to be included. Whereas to add an interactive option for Solaris and SLES, I just copy and paste the boot section in the PXE config file and remove the kernel arg that specifies the automated install info. That's a shame. Has this ever been requested before? If so anyone know the rationale for not doing it? -Kyle > > As for using the same interface that you PXE'd from, check out the > ksdevice=bootif option, described briefly here: > > http://wiki.centos.org/TipsAndTricks/KickStart > > This document from Dell mentions it too: > > http://linux.dell.com/files/whitepapers/nic-enum-whitepaper-v2.pdf > > I don't claim to have figured this out myself, it's set by default on > all my Cobbler PXE profiles. > > /* Wes Hardin */ From lijian.gnu at gmail.com Thu Apr 29 01:45:52 2010 From: lijian.gnu at gmail.com (Jian Lee) Date: Thu, 29 Apr 2010 09:45:52 +0800 Subject: where is the "BaseHandler.__init__" exec ? In-Reply-To: <20100428140449.GR2546@localhost.localdomain> References: <201004281843.29915.lijian.gnu@gmail.com> <20100428140449.GR2546@localhost.localdomain> Message-ID: <201004290945.52511.lijian.gnu@gmail.com> Hi, Chris Thanks for your reply very much. After i debug the code step by step, I'm found the "return cl()" in pykickstart/version.makeVersion() exec the __init__ ! Maybe my description is not clearly, but i'm try to do this. 1. just suppose the DEVEL=F13, use pdb to trace in returnClassForVersion, the v is: --------------------------------- (Pdb) p v ================== 2. and returnClassForVersion return a Class to makeVersion 3. the makeVersion instance f13.F13Handler(), but f13.F13Handler does not have a __init__ func, so exec BaseHandler.__init__ instead. I've learned those knowledge in book, but have not used in mycode. So have puzzle above. Thanks again, for the foolish question. Chris Lumens write? >> where is the "BaseHandler.__init__" exec ? >> >> look at the following code, after func "returnClassForVersion" have executed, >> the "commands" attr of "handler" have fixed by "command-object" mapping already. >> >> but from the line 173 to 177, there is just a imputil.imp.load_module func , >> and the module itself haven't execute a "BaseHandler.__init__" func. >> >> how the "handler" get a fixed attr "commands" ? > >returnClassForVersion just returns the class. It does not return an >instance of that class, so you have to do that yourself. Look at >makeVersion in pykickstart/version.py. That returns an instance of the >class, and that's how the handler gets a commands dict. > >Look at pykickstart/base.py:BaseHandler.__init__. There, >_registerCommands is called which sets up that attr. > >Does that clear things up? > >- Chris > --------------------- Jian Lee [ http://jianlee.ylinux.org ] From larry.brigman at gmail.com Thu Apr 29 04:48:06 2010 From: larry.brigman at gmail.com (Larry Brigman) Date: Wed, 28 Apr 2010 21:48:06 -0700 Subject: Can the NFS media location be passed to anaconda as a kernel arg? In-Reply-To: <4BD87B41.8070908@egenera.com> References: <4BD87B41.8070908@egenera.com> Message-ID: On Wed, Apr 28, 2010 at 11:15 AM, Kyle McDonald wrote: > Hi All, > > I want to setup PXE network installs that are interactive, but I'd like > to limit the amount of questions the user has to enter for each one. The > PXE boot already passes some kernel args to configure the serial > console, and force a vnc install. I'm wondering if I can also supply the > NFS location for the media as a kernel arg so that users don't need to > type it in everytime (It's always the same, and it's long.) If pass a kickstart file location in, this can be handle inside of the kickstart. > > Also I know when doing KickStart installs, you can use ks-device=eth1 to > avoid being prompted for the network device when there is more than one. > Can I use this same option so that the interactive user doesn't have to > know which eth to pick? Are you using pxelinux? If so then you can use 'APPEND 2' to set the boot interface http://syslinux.zytor.com/wiki/index.php/SYSLINUX#IPAPPEND_flag_val_.5BPXELINUX_only.5D > > On a related note, is there a way to have anaconda figure out which > interface the PXE boot booted the machine on and just go with that one? > Right now I need to make 2 different PXE boot config files for the 2 > types of machines I have because one boots on eth0 and the other on eth1. If you use the above and a kickstart file that uses dhcp then the networking stuff should be done. From clumens at redhat.com Thu Apr 29 12:59:56 2010 From: clumens at redhat.com (Chris Lumens) Date: Thu, 29 Apr 2010 08:59:56 -0400 Subject: Can the NFS media location be passed to anaconda as a kernel arg? In-Reply-To: <4BD87B41.8070908@egenera.com> References: <4BD87B41.8070908@egenera.com> Message-ID: <20100429125956.GS2546@localhost.localdomain> > I want to setup PXE network installs that are interactive, but I'd like > to limit the amount of questions the user has to enter for each one. The > PXE boot already passes some kernel args to configure the serial > console, and force a vnc install. I'm wondering if I can also supply the > NFS location for the media as a kernel arg so that users don't need to > type it in everytime (It's always the same, and it's long.) Sure, method= (or stage2=/repo= if you're using F13 or RHEL6 beta). You'd do something like method=nfs:your.server.here:/mnt/wherever Keep in mind if your path plus other parameters is too long, you're going to run up against a size limit and arguments will start getting trimmed or dropped. So if you start to see pretty weird behavior, it might be time to switch to a kickstart file. > Also I know when doing KickStart installs, you can use ks-device=eth1 to > avoid being prompted for the network device when there is more than one. > Can I use this same option so that the interactive user doesn't have to > know which eth to pick? > > On a related note, is there a way to have anaconda figure out which > interface the PXE boot booted the machine on and just go with that one? > Right now I need to make 2 different PXE boot config files for the 2 > types of machines I have because one boots on eth0 and the other on eth1. ksdevice=bootif - Chris From asadiq at k12.com Thu Apr 29 18:22:50 2010 From: asadiq at k12.com (Adnan N. Sadiq) Date: Thu, 29 Apr 2010 14:22:50 -0400 Subject: Kickstart-list Digest, Vol 74, Issue 8 In-Reply-To: Message-ID: help -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of kickstart-list-request at redhat.com Sent: Thursday, April 29, 2010 1:28 PM To: kickstart-list at redhat.com Subject: Kickstart-list Digest, Vol 74, Issue 8 Send Kickstart-list mailing list submissions to kickstart-list at redhat.com To subscribe or unsubscribe via the World Wide Web, visit https://www.redhat.com/mailman/listinfo/kickstart-list or, via email, send a message with subject or body 'help' to kickstart-list-request at redhat.com You can reach the person managing the list at kickstart-list-owner at redhat.com When replying, please edit your Subject line so it is more specific than "Re: Contents of Kickstart-list digest..." Today's Topics: 1. Can the NFS media location be passed to anaconda as a kernel arg? (Kyle McDonald) 2. Re: Can the NFS media location be passed to anaconda as a kernel arg? (Wes Hardin) 3. Re: Can the NFS media location be passed to anaconda as a kernel arg? (Wes Hardin) 4. Re: Can the NFS media location be passed to anaconda as a kernel arg? (Kyle McDonald) 5. Re: where is the "BaseHandler.__init__" exec ? (Jian Lee) 6. Re: Can the NFS media location be passed to anaconda as a kernel arg? (Larry Brigman) 7. Re: Can the NFS media location be passed to anaconda as a kernel arg? (Chris Lumens) ---------------------------------------------------------------------- Message: 1 Date: Wed, 28 Apr 2010 14:15:29 -0400 From: Kyle McDonald To: kickstart-list at redhat.com Subject: Can the NFS media location be passed to anaconda as a kernel arg? Message-ID: <4BD87B41.8070908 at egenera.com> Content-Type: text/plain; charset=ISO-8859-1 Hi All, I want to setup PXE network installs that are interactive, but I'd like to limit the amount of questions the user has to enter for each one. The PXE boot already passes some kernel args to configure the serial console, and force a vnc install. I'm wondering if I can also supply the NFS location for the media as a kernel arg so that users don't need to type it in everytime (It's always the same, and it's long.) Also I know when doing KickStart installs, you can use ks-device=eth1 to avoid being prompted for the network device when there is more than one. Can I use this same option so that the interactive user doesn't have to know which eth to pick? On a related note, is there a way to have anaconda figure out which interface the PXE boot booted the machine on and just go with that one? Right now I need to make 2 different PXE boot config files for the 2 types of machines I have because one boots on eth0 and the other on eth1. Thanks, -Kyle ------------------------------ Message: 2 Date: Wed, 28 Apr 2010 15:36:05 -0500 From: Wes Hardin To: Discussion list about Kickstart Subject: Re: Can the NFS media location be passed to anaconda as a kernel arg? Message-ID: <4BD89C35.3090103 at maxim-ic.com> Content-Type: text/plain; charset="UTF-8"; format=flowed On 04/28/2010 01:15 PM, Kyle McDonald wrote: > I want to setup PXE network installs that are interactive, but I'd like > to limit the amount of questions the user has to enter for each one. The > PXE boot already passes some kernel args to configure the serial > console, and force a vnc install. I'm wondering if I can also supply the > NFS location for the media as a kernel arg so that users don't need to > type it in everytime (It's always the same, and it's long.) > > Also I know when doing KickStart installs, you can use ks-device=eth1 to > avoid being prompted for the network device when there is more than one. > Can I use this same option so that the interactive user doesn't have to > know which eth to pick? > > On a related note, is there a way to have anaconda figure out which > interface the PXE boot booted the machine on and just go with that one? > Right now I need to make 2 different PXE boot config files for the 2 > types of machines I have because one boots on eth0 and the other on eth1. I don't think media location is something you pass in on the kernel line. Instead, what you can pass in is the path to a kickstart file which then answers a lot of those repetitive questions, such as the location of the install tree. The kickstart file can be as complete or as minimal as you wish. I have one that only contains the network location of the install tree and waits for the installer to answer the rest. As for using the same interface that you PXE'd from, check out the ksdevice=bootif option, described briefly here: http://wiki.centos.org/TipsAndTricks/KickStart This document from Dell mentions it too: http://linux.dell.com/files/whitepapers/nic-enum-whitepaper-v2.pdf I don't claim to have figured this out myself, it's set by default on all my Cobbler PXE profiles. /* Wes Hardin */ ------------------------------ Message: 3 Date: Wed, 28 Apr 2010 15:47:39 -0500 From: Wes Hardin To: Discussion list about Kickstart Subject: Re: Can the NFS media location be passed to anaconda as a kernel arg? Message-ID: <4BD89EEB.1030705 at maxim-ic.com> Content-Type: text/plain; charset="UTF-8"; format=flowed On 04/28/2010 03:36 PM, Wes Hardin wrote: > On 04/28/2010 01:15 PM, Kyle McDonald wrote: >> I want to setup PXE network installs that are interactive, but I'd like >> to limit the amount of questions the user has to enter for each one. The >> PXE boot already passes some kernel args to configure the serial >> console, and force a vnc install. I'm wondering if I can also supply the >> NFS location for the media as a kernel arg so that users don't need to >> type it in everytime (It's always the same, and it's long.) >> >> Also I know when doing KickStart installs, you can use ks-device=eth1 to >> avoid being prompted for the network device when there is more than one. >> Can I use this same option so that the interactive user doesn't have to >> know which eth to pick? >> >> On a related note, is there a way to have anaconda figure out which >> interface the PXE boot booted the machine on and just go with that one? >> Right now I need to make 2 different PXE boot config files for the 2 >> types of machines I have because one boots on eth0 and the other on eth1. > > I don't think media location is something you pass in on the kernel line. > Instead, what you can pass in is the path to a kickstart file which then answers > a lot of those repetitive questions, such as the location of the install tree. > The kickstart file can be as complete or as minimal as you wish. I have one > that only contains the network location of the install tree and waits for the > installer to answer the rest. > > As for using the same interface that you PXE'd from, check out the > ksdevice=bootif option, described briefly here: > > http://wiki.centos.org/TipsAndTricks/KickStart > > This document from Dell mentions it too: > > http://linux.dell.com/files/whitepapers/nic-enum-whitepaper-v2.pdf > > I don't claim to have figured this out myself, it's set by default on all my > Cobbler PXE profiles. And just as a note, I don't think this works on RHEL4. I had a workstation last week with 2 interfaces that I ended up just disabling the second interface to get it to work sanely. Here's the note on the change from the Cobbler folks, which includes a link to the relevant Fedora 9 documentation: https://fedorahosted.org/pipermail/cobbler/2008-September/000824.html -- /* Wes Hardin */ ------------------------------ Message: 4 Date: Wed, 28 Apr 2010 17:05:42 -0400 From: Kyle McDonald To: Wes Hardin Cc: Discussion list about Kickstart Subject: Re: Can the NFS media location be passed to anaconda as a kernel arg? Message-ID: <4BD8A326.7090604 at egenera.com> Content-Type: text/plain; charset=UTF-8 On 4/28/2010 4:36 PM, Wes Hardin wrote: > On 04/28/2010 01:15 PM, Kyle McDonald wrote: >> I want to setup PXE network installs that are interactive, but I'd like >> to limit the amount of questions the user has to enter for each one. The >> PXE boot already passes some kernel args to configure the serial >> console, and force a vnc install. I'm wondering if I can also supply the >> NFS location for the media as a kernel arg so that users don't need to >> type it in everytime (It's always the same, and it's long.) >> >> Also I know when doing KickStart installs, you can use ks-device=eth1 to >> avoid being prompted for the network device when there is more than one. >> Can I use this same option so that the interactive user doesn't have to >> know which eth to pick? >> >> On a related note, is there a way to have anaconda figure out which >> interface the PXE boot booted the machine on and just go with that one? >> Right now I need to make 2 different PXE boot config files for the 2 >> types of machines I have because one boots on eth0 and the other on >> eth1. > > I don't think media location is something you pass in on the kernel > line. Instead, what you can pass in is the path to a kickstart file > which then answers a lot of those repetitive questions, such as the > location of the install tree. The kickstart file can be as complete or > as minimal as you wish. I have one that only contains the network > location of the install tree and waits for the installer to answer the > rest. That blows. I already have a PXE config file for each release of RHEL SLES and Solaris I need to install. These are required to direct the network boot to the right release. For SLES and Solaris, I can put the media location in this PXE config file along with the install config file (AutoYast or JumpStart), and I can get away with just 1 install config file for all my machines (I use the begin and finish scripts to tweak the profile based on what release was booted so 1 file really helps me out since when the I do need to make a code update I only need to edit it once.) But for RHEL (and CENTOS) I have to have 2 (32 and 64 bit) KS config files for each release that only differ by the 'nfs ....' line. Now If I want to have an interactive install where the media location is already set, I need to go and make 2 more KS files for every release? That's # releases * 4 files to update when and if the rest of the file needs to change. And the 'nfs ...' line is one of the things that the %pre can't generate when it writes out the file to be included. Whereas to add an interactive option for Solaris and SLES, I just copy and paste the boot section in the PXE config file and remove the kernel arg that specifies the automated install info. That's a shame. Has this ever been requested before? If so anyone know the rationale for not doing it? -Kyle > > As for using the same interface that you PXE'd from, check out the > ksdevice=bootif option, described briefly here: > > http://wiki.centos.org/TipsAndTricks/KickStart > > This document from Dell mentions it too: > > http://linux.dell.com/files/whitepapers/nic-enum-whitepaper-v2.pdf > > I don't claim to have figured this out myself, it's set by default on > all my Cobbler PXE profiles. > > /* Wes Hardin */ ------------------------------ Message: 5 Date: Thu, 29 Apr 2010 09:45:52 +0800 From: Jian Lee To: Chris Lumens Cc: kickstart-list at redhat.com Subject: Re: where is the "BaseHandler.__init__" exec ? Message-ID: <201004290945.52511.lijian.gnu at gmail.com> Content-Type: Text/Plain; charset="utf-8" Hi, Chris Thanks for your reply very much. After i debug the code step by step, I'm found the "return cl()" in pykickstart/version.makeVersion() exec the __init__ ! Maybe my description is not clearly, but i'm try to do this. 1. just suppose the DEVEL=F13, use pdb to trace in returnClassForVersion, the v is: --------------------------------- (Pdb) p v ================== 2. and returnClassForVersion return a Class to makeVersion 3. the makeVersion instance f13.F13Handler(), but f13.F13Handler does not have a __init__ func, so exec BaseHandler.__init__ instead. I've learned those knowledge in book, but have not used in mycode. So have puzzle above. Thanks again, for the foolish question. Chris Lumens write? >> where is the "BaseHandler.__init__" exec ? >> >> look at the following code, after func "returnClassForVersion" have executed, >> the "commands" attr of "handler" have fixed by "command-object" mapping already. >> >> but from the line 173 to 177, there is just a imputil.imp.load_module func , >> and the module itself haven't execute a "BaseHandler.__init__" func. >> >> how the "handler" get a fixed attr "commands" ? > >returnClassForVersion just returns the class. It does not return an >instance of that class, so you have to do that yourself. Look at >makeVersion in pykickstart/version.py. That returns an instance of the >class, and that's how the handler gets a commands dict. > >Look at pykickstart/base.py:BaseHandler.__init__. There, >_registerCommands is called which sets up that attr. > >Does that clear things up? > >- Chris > --------------------- Jian Lee [ http://jianlee.ylinux.org ] ------------------------------ Message: 6 Date: Wed, 28 Apr 2010 21:48:06 -0700 From: Larry Brigman To: Discussion list about Kickstart Subject: Re: Can the NFS media location be passed to anaconda as a kernel arg? Message-ID: Content-Type: text/plain; charset=ISO-8859-1 On Wed, Apr 28, 2010 at 11:15 AM, Kyle McDonald wrote: > Hi All, > > I want to setup PXE network installs that are interactive, but I'd like > to limit the amount of questions the user has to enter for each one. The > PXE boot already passes some kernel args to configure the serial > console, and force a vnc install. I'm wondering if I can also supply the > NFS location for the media as a kernel arg so that users don't need to > type it in everytime (It's always the same, and it's long.) If pass a kickstart file location in, this can be handle inside of the kickstart. > > Also I know when doing KickStart installs, you can use ks-device=eth1 to > avoid being prompted for the network device when there is more than one. > Can I use this same option so that the interactive user doesn't have to > know which eth to pick? Are you using pxelinux? If so then you can use 'APPEND 2' to set the boot interface http://syslinux.zytor.com/wiki/index.php/SYSLINUX#IPAPPEND_flag_val_.5BP XELINUX_only.5D > > On a related note, is there a way to have anaconda figure out which > interface the PXE boot booted the machine on and just go with that one? > Right now I need to make 2 different PXE boot config files for the 2 > types of machines I have because one boots on eth0 and the other on eth1. If you use the above and a kickstart file that uses dhcp then the networking stuff should be done. ------------------------------ Message: 7 Date: Thu, 29 Apr 2010 08:59:56 -0400 From: Chris Lumens To: kickstart-list at redhat.com Subject: Re: Can the NFS media location be passed to anaconda as a kernel arg? Message-ID: <20100429125956.GS2546 at localhost.localdomain> Content-Type: text/plain; charset=us-ascii > I want to setup PXE network installs that are interactive, but I'd like > to limit the amount of questions the user has to enter for each one. The > PXE boot already passes some kernel args to configure the serial > console, and force a vnc install. I'm wondering if I can also supply the > NFS location for the media as a kernel arg so that users don't need to > type it in everytime (It's always the same, and it's long.) Sure, method= (or stage2=/repo= if you're using F13 or RHEL6 beta). You'd do something like method=nfs:your.server.here:/mnt/wherever Keep in mind if your path plus other parameters is too long, you're going to run up against a size limit and arguments will start getting trimmed or dropped. So if you start to see pretty weird behavior, it might be time to switch to a kickstart file. > Also I know when doing KickStart installs, you can use ks-device=eth1 to > avoid being prompted for the network device when there is more than one. > Can I use this same option so that the interactive user doesn't have to > know which eth to pick? > > On a related note, is there a way to have anaconda figure out which > interface the PXE boot booted the machine on and just go with that one? > Right now I need to make 2 different PXE boot config files for the 2 > types of machines I have because one boots on eth0 and the other on eth1. ksdevice=bootif - Chris ------------------------------ _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list End of Kickstart-list Digest, Vol 74, Issue 8 ********************************************* From kmcdonald at egenera.com Thu Apr 29 18:26:21 2010 From: kmcdonald at egenera.com (Kyle McDonald) Date: Thu, 29 Apr 2010 14:26:21 -0400 Subject: Can the NFS media location be passed to anaconda as a kernel arg? In-Reply-To: <20100429125956.GS2546@localhost.localdomain> References: <4BD87B41.8070908@egenera.com> <20100429125956.GS2546@localhost.localdomain> Message-ID: <4BD9CF4D.9010206@egenera.com> On 4/29/2010 8:59 AM, Chris Lumens wrote: >> I want to setup PXE network installs that are interactive, but I'd like >> to limit the amount of questions the user has to enter for each one. The >> PXE boot already passes some kernel args to configure the serial >> console, and force a vnc install. I'm wondering if I can also supply the >> NFS location for the media as a kernel arg so that users don't need to >> type it in everytime (It's always the same, and it's long.) >> > Sure, method= (or stage2=/repo= if you're using F13 or RHEL6 beta). > > You'd do something like method=nfs:your.server.here:/mnt/wherever > > Cool. Thanks! > Keep in mind if your path plus other parameters is too long, you're > going to run up against a size limit and arguments will start getting > trimmed or dropped. So if you start to see pretty weird behavior, it > might be time to switch to a kickstart file. > I'm aware of that with SLES, but luckily I haven't hit that problem yet. >> Also I know when doing KickStart installs, you can use ks-device=eth1 to >> avoid being prompted for the network device when there is more than one. >> Can I use this same option so that the interactive user doesn't have to >> know which eth to pick? >> >> On a related note, is there a way to have anaconda figure out which >> interface the PXE boot booted the machine on and just go with that one? >> Right now I need to make 2 different PXE boot config files for the 2 >> types of machines I have because one boots on eth0 and the other on eth1. >> > ksdevice=bootif > > That's cool too. What release did that appear in? Now if GRUB or PXElinux could only read the serial console redirection settings from the BIOS, and pass them to the kernel, I could cut the number of PXE config files I have in half (half ttyS0 at 115200, the others at ttyS1 at 19200 - Ugh!) -Kyle > - Chris > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list >