From itamar at ispbrasil.com.br Wed Nov 9 21:25:54 2011 From: itamar at ispbrasil.com.br (Itamar Reis Peixoto) Date: Wed, 9 Nov 2011 19:25:54 -0200 Subject: setting up proxy during install Message-ID: how can I setup a system proxy during install ? something like that http://askubuntu.com/questions/63075/is-there-a-way-to-create-a-shortcut-to-proxy-preferences-in-the-launcher -- ------------ Itamar Reis Peixoto msn, google talk: itamar at ispbrasil.com.br +55 11 4063 5033 (FIXO SP) +55 34 9158 9329 (TIM) +55 34 8806 3989 (OI) +55 34 3221 8599 (FIXO MG) From co55-sy1t at dea.spamcon.org Sun Nov 13 19:53:03 2011 From: co55-sy1t at dea.spamcon.org (whitivery) Date: Sun, 13 Nov 2011 11:53:03 -0800 Subject: Driver disk loading via NFS fails on Cisco switch, OK on Dell Message-ID: I have a working Cobbler server, using it to load units with CentOS 5.7. It works fine with a simple Dell 2716 switch connecting the server and targets. Also with a LinkSys RV016. Substituting a Cisco 2960 switch, it fails. It network boots, shows the PXE menu, I choose an entry, it gets a DHCP address on its eth0, and starts loading including getting its kickstart via HTTP. But when it goes to load the driver disk image via NFS, it hangs. The Anaconda Alt-F4 screen shows it not seeing the NFS server; after 2 minutes of trying, it times out and the Alt-F4 screen shows an error dump. Tried the switch with and without RST (rapid spanning tree) active, no difference. The switch seems to work OK for normal network operations outside the kickstart environment, for what we use. Any idea what it might be that could be interfering with the NFS disk driver loading step? From root at nachtmaus.us Mon Nov 14 01:20:21 2011 From: root at nachtmaus.us (david klein) Date: Sun, 13 Nov 2011 19:20:21 -0600 Subject: Driver disk loading via NFS fails on Cisco switch, OK on Dell In-Reply-To: References: Message-ID: Check to see if the client, switch and up-stream router agree on maximum frame-size (i.e. if one is trying to do jumbo-frames and the other can't, it can cause troubles). Make sure the switch-port that the client is using is setup as an access-port, not a trunk. Set the client's port for spanning-tree portfast and turn on BPDU-guard. On Sun, Nov 13, 2011 at 1:53 PM, whitivery wrote: > I have a working Cobbler server, using it to load units with > CentOS 5.7. It works fine with a simple Dell 2716 switch > connecting the server and targets. Also with a LinkSys RV016. > > Substituting a Cisco 2960 switch, it fails. It network boots, > shows the PXE menu, I choose an entry, it gets a DHCP address on > its eth0, and starts loading including getting its kickstart via > HTTP. But when it goes to load the driver disk image via NFS, it > hangs. The Anaconda Alt-F4 screen shows it not seeing the NFS > server; after 2 minutes of trying, it times out and the Alt-F4 > screen shows an error dump. > > Tried the switch with and without RST (rapid spanning tree) > active, no difference. > > The switch seems to work OK for normal network operations outside > the kickstart environment, for what we use. > > Any idea what it might be that could be interfering with the NFS > disk driver loading step? > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- david t. klein Cisco Certified Network Associate (CSCO11281885) Linux Professional Institute Certification (LPI000165615) Redhat Certified Engineer (805009745938860) Quis custodiet ipsos custodes? -------------- next part -------------- An HTML attachment was scrubbed... URL: From co55-sy1t at dea.spamcon.org Tue Nov 15 08:45:08 2011 From: co55-sy1t at dea.spamcon.org (whitivery) Date: Tue, 15 Nov 2011 00:45:08 -0800 Subject: Driver disk loading via NFS fails on Cisco switch, OK on Dell References: Message-ID: <7594c75jk53025t5mpvuva5d21v3j83mev@4ax.com> david klein wrote: >Check to see if the client, switch and up-stream router agree on maximum >frame-size (i.e. if one is trying to do jumbo-frames and the other can't, >it can cause troubles). Make sure the switch-port that the client is using >is setup as an access-port, not a trunk. Set the client's port for >spanning-tree portfast and turn on BPDU-guard. Thank you for the response. Had already tried spanning-tree portfast before turning off spanning-tree entirely, no difference. I will look into the other things you mention. Updated observations: - Problem happens exactly the same with a switch at the opposite end of the complexity spectrum - Netgear 5 port 10/100Mbps Switch FS605 v3. - I noticed that on the good switch, you see two eth0 dialogs flash up before it proceeds; on a bad switch, only see one eth0 dialog then the empty blue screen. - If while it's in the failure state with the empty blue screen, I move the target's network connection to the good switch, it immediately takes off and continues the kickstart process properly. >On Sun, Nov 13, 2011 at 1:53 PM, whitivery wrote: > >> I have a working Cobbler server, using it to load units with >> CentOS 5.7. It works fine with a simple Dell 2716 switch >> connecting the server and targets. Also with a LinkSys RV016. >> >> Substituting a Cisco 2960 switch, it fails. It network boots, >> shows the PXE menu, I choose an entry, it gets a DHCP address on >> its eth0, and starts loading including getting its kickstart via >> HTTP. But when it goes to load the driver disk image via NFS, it >> hangs. The Anaconda Alt-F4 screen shows it not seeing the NFS >> server; after 2 minutes of trying, it times out and the Alt-F4 >> screen shows an error dump. >> >> Tried the switch with and without RST (rapid spanning tree) >> active, no difference. >> >> The switch seems to work OK for normal network operations outside >> the kickstart environment, for what we use. >> >> Any idea what it might be that could be interfering with the NFS >> disk driver loading step? >> >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list >> From gerhardus.geldenhuis at gmail.com Mon Nov 28 12:56:41 2011 From: gerhardus.geldenhuis at gmail.com (Gerhardus Geldenhuis) Date: Mon, 28 Nov 2011 12:56:41 +0000 Subject: Alligning partitions Message-ID: Hi I am building Red Hat 5.4 and 5.5 boxes on VMWare and some on HP blades with SAN attached storage. I need to do partition alignment during build which I understand is done automatically in Red Hat 6's version of anaconda and later but not yet in 5.4, 5.5. So a number of questions: * I could not find an option for part in the docs ( http://fedoraproject.org/wiki/Anaconda/Kickstart#part_or_partition) that easily allow me to set a sector offset. I have found a number of ways to do this but was wondering if there is an hidden option to do this using part. Currently I am using fdisk to create a partition starting at sector 64 rather than 63 for SAN storage and doing the same for VMWare boxes. * How can I easily debug my kickstart? I currently change the kickstart and then reboot the physical server but that is very painful as it has to go through all of the bios checks etc and then sometime I forget to press F12... I tried adding a sleep 100000 to my pre scripts which works and allows me to run some commands on the command line but I am not sure how to "reparse" the kickstart file to test whether new settings works. * Also rather than use clever ways of getting a list of drives or anything similar... is there predefined procedures/variables that I can use, are they available in all scripting languages (bash, python) and where/how could I easily get a list of available variables for scripting. If I need to rtfm, the please point me in the right direction. Best Regards -- Gerhardus Geldenhuis -------------- next part -------------- An HTML attachment was scrubbed... URL: From klaus.arndt at verigy.com Tue Nov 29 07:31:59 2011 From: klaus.arndt at verigy.com (Arndt, Klaus) Date: Tue, 29 Nov 2011 07:31:59 +0000 Subject: Alligning partitions In-Reply-To: References: Message-ID: Hello, For RHEL5 partition alignment we use "parted" insted "fdisk" in the kickstart scripts. Then you need to calculate the Start Sectors of your partition, make sure they are exact divisible with 8 (8x512Byte = 4096Byte) # SS_BOOT= Start Sector /boot partition # SS_ROOT= Start Sector / partition # SS_SWAP= Start Sector swap partition # LASTSECTOR= Last usable Sector (exact divisible with 8, -1) parted -s -- ${DISC} mktable msdos parted -s -- ${DISC} mkpart primary ext2 ${SS_BOOT}s $(( ${SS_ROOT} - 1 ))s parted -s -- ${DISC} mkpart primary ext2 ${SS_ROOT}s $(( ${SS_SWAP} - 1 ))s parted -s -- ${DISC} mkpartfs primary linux-swap ${SS_SWAP}s ${LASTSECTOR}s Regards From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Gerhardus Geldenhuis Sent: Montag, 28. November 2011 13:57 To: kickstart-list at redhat.com Subject: Alligning partitions Hi I am building Red Hat 5.4 and 5.5 boxes on VMWare and some on HP blades with SAN attached storage. I need to do partition alignment during build which I understand is done automatically in Red Hat 6's version of anaconda and later but not yet in 5.4, 5.5. So a number of questions: * I could not find an option for part in the docs (http://fedoraproject.org/wiki/Anaconda/Kickstart#part_or_partition) that easily allow me to set a sector offset. I have found a number of ways to do this but was wondering if there is an hidden option to do this using part. Currently I am using fdisk to create a partition starting at sector 64 rather than 63 for SAN storage and doing the same for VMWare boxes. * How can I easily debug my kickstart? I currently change the kickstart and then reboot the physical server but that is very painful as it has to go through all of the bios checks etc and then sometime I forget to press F12... I tried adding a sleep 100000 to my pre scripts which works and allows me to run some commands on the command line but I am not sure how to "reparse" the kickstart file to test whether new settings works. * Also rather than use clever ways of getting a list of drives or anything similar... is there predefined procedures/variables that I can use, are they available in all scripting languages (bash, python) and where/how could I easily get a list of available variables for scripting. If I need to rtfm, the please point me in the right direction. Best Regards -- Gerhardus Geldenhuis -------------- next part -------------- An HTML attachment was scrubbed... URL: From petro at bounty.org Tue Nov 29 09:11:51 2011 From: petro at bounty.org (Petro) Date: Tue, 29 Nov 2011 18:41:51 +0930 Subject: Alligning partitions In-Reply-To: References: Message-ID: <4ED4A1D7.8050704@bounty.org> On 11/28/11 22:26 PM, Gerhardus Geldenhuis wrote: > Hi > I am building Red Hat 5.4 and 5.5 boxes on VMWare and some on HP > blades with SAN attached storage. I need to do > partition alignment during build which I understand is done > automatically in Red Hat 6's version of anaconda and later but not yet > in 5.4, 5.5. So a number of questions: > * I could not find an option for part in the > docs (http://fedoraproject.org/wiki/Anaconda/Kickstart#part_or_partition) > that easily allow me to set a sector offset. I have found a number of > ways to do this but was wondering if there is an hidden option to do > this using part. Currently I am using fdisk to create a partition > starting at sector 64 rather than 63 for SAN storage and doing the > same for VMWare boxes. I *thought* (check with VMware) that after Vsphere 3 you wouldn't need to align the partitions any more--but the last time I checked was in 2009, so I could be seriously wrong. Or you could be running V3, in which case... > * How can I easily debug my kickstart? I currently change the > kickstart and then reboot the physical server but that is > very painful as it has to go through all of the bios checks etc and > then sometime I forget to press F12... I tried adding a sleep 100000 > to my pre scripts which works and allows me to run some commands on > the command line but I am not sure how to "reparse" the kickstart file > to test whether new settings works. I usually use VMware Workstation/Player to debug my kickstart scripts, with lots of redirecting to different log files. This lets me sit in my somewhat comfy chair in a slightly quieter machine room (my desk is on a raised floor, but most of the servers are in the other room) and watch the installs while I do other things. Then when I've got it mostly right I run it on the real hardware, which would (eventually) be a known transform. > * Also rather than use clever ways of getting a list of drives or > anything similar... is there predefined procedures/variables that I > can use, are they available in all scripting languages (bash, python) > and where/how could I easily get a list of available variables for > scripting. I suspect if there was a clever call for it in kickstart that underneath it would use a "clever" way and be not so portable to new hardware. This is actually a more difficult problem than you'd think at first, what with iSCSI, FCOE, SCSI, SATA, IDE, whatever HP does to get /dev/cciss/c0d0 stuff etc. Would be nice though. Regards, Petro. L at b R at t. -------------- next part -------------- An HTML attachment was scrubbed... URL: From petro at cpetro.us Wed Nov 30 08:50:34 2011 From: petro at cpetro.us (Petro) Date: Wed, 30 Nov 2011 18:20:34 +0930 Subject: Aligning partitions In-Reply-To: References: Message-ID: <4ED5EE5A.10902@cpetro.us> On 11/30/11 02:30 AM, I wrote from a different account: > Re: Alligning partitions > > >> Hi >> I am building Red Hat 5.4 and 5.5 boxes on VMWare and some on HP >> blades with SAN attached storage. I need to do >> partition alignment during build which I understand is done >> automatically in Red Hat 6's version of anaconda and later but not >> yet in 5.4, 5.5. So a number of questions: >> * I could not find an option for part in the >> docs (http://fedoraproject.org/wiki/Anaconda/Kickstart#part_or_partition) >> that easily allow me to set a sector offset. I have found a number of >> ways to do this but was wondering if there is an hidden option to do >> this using part. Currently I am using fdisk to create a partition >> starting at sector 64 rather than 63 for SAN storage and doing the >> same for VMWare boxes. > > I *thought* (check with VMware) that after Vsphere 3 you wouldn't need > to align the partitions any more--but the last time I checked was in > 2009, so I could be seriously wrong. Or you could be running V3, in > which case... I was wrong about this, at least in 4.0. >> * How can I easily debug my kickstart? I currently change the >> kickstart and then reboot the physical server but that is >> very painful as it has to go through all of the bios checks etc and >> then sometime I forget to press F12... I tried adding a sleep 100000 >> to my pre scripts which works and allows me to run some commands on >> the command line but I am not sure how to "reparse" the kickstart >> file to test whether new settings works. > > I usually use VMware Workstation/Player to debug my kickstart scripts, > with lots of redirecting to different log files. This lets me sit in > my somewhat comfy chair in a slightly quieter machine room (my desk is > on a raised floor, but most of the servers are in the other room) and > watch the installs while I do other things. Then when I've got it > mostly right I run it on the real hardware, which would (eventually) > be a known transform. One of my co-workers pointed me to an rpm called "pykickstart" that contains ksvalidator. It seems to work as it gave me the same errors on my ks.cfg as anaconda did. I still don't understand why --resolution and --depth don't take any options in 6.1, when they don't make sense without options. Regards, Petro. :wq -------------- next part -------------- An HTML attachment was scrubbed... URL: From co55-sy1t at dea.spamcon.org Wed Nov 30 09:05:39 2011 From: co55-sy1t at dea.spamcon.org (whitivery) Date: Wed, 30 Nov 2011 01:05:39 -0800 Subject: Driver disk loading via NFS fails on Cisco switch, OK on Dell References: <7594c75jk53025t5mpvuva5d21v3j83mev@4ax.com> Message-ID: whitivery wrote: >david klein wrote: > >>Check to see if the client, switch and up-stream router agree on maximum >>frame-size (i.e. if one is trying to do jumbo-frames and the other can't, >>it can cause troubles). Make sure the switch-port that the client is using >>is setup as an access-port, not a trunk. Set the client's port for >>spanning-tree portfast and turn on BPDU-guard. > >Thank you for the response. Had already tried spanning-tree portfast >before turning off spanning-tree entirely, no difference. I will look >into the other things you mention. > >Updated observations: > >- Problem happens exactly the same with a switch at the opposite > end of the complexity spectrum - Netgear 5 port 10/100Mbps > Switch FS605 v3. > >- I noticed that on the good switch, you see two eth0 dialogs > flash up before it proceeds; on a bad switch, only see one eth0 > dialog then the empty blue screen. > >- If while it's in the failure state with the empty blue screen, > I move the target's network connection to the good switch, it > immediately takes off and continues the kickstart process > properly. Answering myself here: problem is solved. I was using NFS to load the driver disk. I changed to HTTP, no more problem, works with Netgear and Cisco just fine. > >>On Sun, Nov 13, 2011 at 1:53 PM, whitivery wrote: >> >>> I have a working Cobbler server, using it to load units with >>> CentOS 5.7. It works fine with a simple Dell 2716 switch >>> connecting the server and targets. Also with a LinkSys RV016. >>> >>> Substituting a Cisco 2960 switch, it fails. It network boots, >>> shows the PXE menu, I choose an entry, it gets a DHCP address on >>> its eth0, and starts loading including getting its kickstart via >>> HTTP. But when it goes to load the driver disk image via NFS, it >>> hangs. The Anaconda Alt-F4 screen shows it not seeing the NFS >>> server; after 2 minutes of trying, it times out and the Alt-F4 >>> screen shows an error dump. >>> >>> Tried the switch with and without RST (rapid spanning tree) >>> active, no difference. >>> >>> The switch seems to work OK for normal network operations outside >>> the kickstart environment, for what we use. >>> >>> Any idea what it might be that could be interfering with the NFS >>> disk driver loading step? >>> >>> >>> _______________________________________________ >>> Kickstart-list mailing list >>> Kickstart-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/kickstart-list >>> From CallahanT at tessco.com Wed Nov 30 10:52:01 2011 From: CallahanT at tessco.com (Callahan, Tom) Date: Wed, 30 Nov 2011 10:52:01 +0000 Subject: Aligning partitions In-Reply-To: <4ED5EE5A.10902@cpetro.us> Message-ID: You might want to check out this tool for aligning: http://nickapedia.com/2011/11/03/straighten-up-with-a-new-uber-tool-presenting-uberalign/ -- Tom Callahan Technology, Development and Services (TDS) 410-229-1361 Tel 410-229-1512 Fax CallahanT at TESSCO.com Visit TESSCO.com TESSCO Your Total Source? for Everything Wireless Network Infrastructure Equipment | Mobile Devices & Accessories | Installation, Test, Equipment & Supplies From: Petro > Reply-To: Discussion list about Kickstart > Date: Wed, 30 Nov 2011 18:20:34 +0930 To: > Subject: Re:Aligning partitions On 11/30/11 02:30 AM, I wrote from a different account: Re: Alligning partitions Hi I am building Red Hat 5.4 and 5.5 boxes on VMWare and some on HP blades with SAN attached storage. I need to do partition alignment during build which I understand is done automatically in Red Hat 6's version of anaconda and later but not yet in 5.4, 5.5. So a number of questions: * I could not find an option for part in the docs (http://fedoraproject.org/wiki/Anaconda/Kickstart#part_or_partition) that easily allow me to set a sector offset. I have found a number of ways to do this but was wondering if there is an hidden option to do this using part. Currently I am using fdisk to create a partition starting at sector 64 rather than 63 for SAN storage and doing the same for VMWare boxes. I *thought* (check with VMware) that after Vsphere 3 you wouldn't need to align the partitions any more--but the last time I checked was in 2009, so I could be seriously wrong. Or you could be running V3, in which case... I was wrong about this, at least in 4.0. * How can I easily debug my kickstart? I currently change the kickstart and then reboot the physical server but that is very painful as it has to go through all of the bios checks etc and then sometime I forget to press F12... I tried adding a sleep 100000 to my pre scripts which works and allows me to run some commands on the command line but I am not sure how to "reparse" the kickstart file to test whether new settings works. I usually use VMware Workstation/Player to debug my kickstart scripts, with lots of redirecting to different log files. This lets me sit in my somewhat comfy chair in a slightly quieter machine room (my desk is on a raised floor, but most of the servers are in the other room) and watch the installs while I do other things. Then when I've got it mostly right I run it on the real hardware, which would (eventually) be a known transform. One of my co-workers pointed me to an rpm called "pykickstart" that contains ksvalidator. It seems to work as it gave me the same errors on my ks.cfg as anaconda did. I still don't understand why --resolution and --depth don't take any options in 6.1, when they don't make sense without options. Regards, Petro. :wq _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list -------------- next part -------------- An HTML attachment was scrubbed... URL: From clumens at redhat.com Wed Nov 30 15:52:24 2011 From: clumens at redhat.com (Chris Lumens) Date: Wed, 30 Nov 2011 10:52:24 -0500 Subject: Aligning partitions In-Reply-To: <4ED5EE5A.10902@cpetro.us> References: <4ED5EE5A.10902@cpetro.us> Message-ID: <20111130155223.GD23005@exeter.usersys.redhat.com> > I still don't understand why --resolution and --depth don't take any > options in 6.1, when they don't make sense without options. These are both deprecated options so the message about not taking any argument is just kind of dumb error reporting. - Chris From petro at cpetro.us Wed Nov 30 16:07:21 2011 From: petro at cpetro.us (Petro) Date: Thu, 01 Dec 2011 01:37:21 +0930 Subject: Kickstart-list Digest, Vol 93, Issue 6 In-Reply-To: References: Message-ID: <4ED654B9.7040001@cpetro.us> Chris Lumens wrote: > > > > >> > I still don't understand why --resolution and --depth don't take any >> > options in 6.1, when they don't make sense without options. > These are both deprecated options so the message about not taking any > argument is just kind of dumb error reporting. Damn. I must have been looking at hold documentation because I saw it in there, but checking online right now it's not there. Oops. Regards, Petro. :wq -------------- next part -------------- An HTML attachment was scrubbed... URL: