You can subscribe to this list here.
2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(1) |
Aug
|
Sep
(1) |
Oct
(1) |
Nov
|
Dec
(9) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2003 |
Jan
(10) |
Feb
(2) |
Mar
(3) |
Apr
(2) |
May
(3) |
Jun
(50) |
Jul
(41) |
Aug
(76) |
Sep
(125) |
Oct
(83) |
Nov
(89) |
Dec
(32) |
2004 |
Jan
(78) |
Feb
(82) |
Mar
(129) |
Apr
(72) |
May
(110) |
Jun
(70) |
Jul
(76) |
Aug
(64) |
Sep
(41) |
Oct
(51) |
Nov
(36) |
Dec
(34) |
2005 |
Jan
(77) |
Feb
(57) |
Mar
(38) |
Apr
(26) |
May
(55) |
Jun
(68) |
Jul
(99) |
Aug
(80) |
Sep
(82) |
Oct
(141) |
Nov
(137) |
Dec
(51) |
2006 |
Jan
(133) |
Feb
(33) |
Mar
(156) |
Apr
(104) |
May
(88) |
Jun
(128) |
Jul
(110) |
Aug
(129) |
Sep
(74) |
Oct
(144) |
Nov
(127) |
Dec
(51) |
2007 |
Jan
(140) |
Feb
(140) |
Mar
(100) |
Apr
(42) |
May
(34) |
Jun
(13) |
Jul
(46) |
Aug
(46) |
Sep
(46) |
Oct
(42) |
Nov
(125) |
Dec
(23) |
2008 |
Jan
(15) |
Feb
(77) |
Mar
(57) |
Apr
(68) |
May
(48) |
Jun
(22) |
Jul
(54) |
Aug
(83) |
Sep
(41) |
Oct
(8) |
Nov
(37) |
Dec
(47) |
2009 |
Jan
(13) |
Feb
(11) |
Mar
(14) |
Apr
(55) |
May
(31) |
Jun
(18) |
Jul
(27) |
Aug
(17) |
Sep
(13) |
Oct
(39) |
Nov
(21) |
Dec
(16) |
2010 |
Jan
(10) |
Feb
(5) |
Mar
(15) |
Apr
(4) |
May
(9) |
Jun
(7) |
Jul
(20) |
Aug
(35) |
Sep
(37) |
Oct
(6) |
Nov
(12) |
Dec
(7) |
2011 |
Jan
(9) |
Feb
(6) |
Mar
(9) |
Apr
(78) |
May
(13) |
Jun
(6) |
Jul
(29) |
Aug
(1) |
Sep
|
Oct
(3) |
Nov
(5) |
Dec
(12) |
2012 |
Jan
(1) |
Feb
(24) |
Mar
(12) |
Apr
(7) |
May
(12) |
Jun
(19) |
Jul
(8) |
Aug
(2) |
Sep
|
Oct
|
Nov
(35) |
Dec
(30) |
2013 |
Jan
(11) |
Feb
(13) |
Mar
(15) |
Apr
(18) |
May
(4) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(5) |
Dec
(2) |
2014 |
Jan
(23) |
Feb
(4) |
Mar
(11) |
Apr
(1) |
May
|
Jun
(1) |
Jul
(6) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2015 |
Jan
(6) |
Feb
|
Mar
|
Apr
(1) |
May
(5) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2017 |
Jan
|
Feb
|
Mar
|
Apr
(6) |
May
(2) |
Jun
(7) |
Jul
|
Aug
|
Sep
(2) |
Oct
|
Nov
|
Dec
|
2018 |
Jan
|
Feb
|
Mar
(3) |
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(2) |
Nov
(2) |
Dec
|
2019 |
Jan
|
Feb
|
Mar
(6) |
Apr
(1) |
May
(1) |
Jun
|
Jul
(4) |
Aug
(1) |
Sep
|
Oct
|
Nov
|
Dec
(1) |
2020 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(1) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(7) |
Nov
(6) |
Dec
|
2021 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(1) |
Oct
|
Nov
|
Dec
|
2022 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(3) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Büchmann H. <en...@bu...> - 2012-07-27 11:38:02
|
Dear Marianne, Sorry for the late response - summer and all that. I have encountered something similar, and if I remember correctly, then si_prepareclient is rather picky with the namin of the initrd file (to be packed with the kernel for later installs). si_prepareclient looks for matching file names in /boot/, but it is unaware of the newer forms "initramfs*", and only looks for "initrd*". (Can't remember if you also needthe output from uname -r in there.) My personal hack is simply to copy (or hard link) the "real" initramfs to something, which will be recognized, e.g. like: cp -av initramfs-$(uname -r).img initrd-to-sisuite-$(uname -r).img (Also, I have to remove any /boot/initrd-$(uname -r)kdump.img, but you do not seem to have such). Best, Bjarne PS: The actual (wrong) checking for si_prepareclient goes on in the sub "is_initrd" in UseYourOwnKernel.pm. On 18-07-2012 8:13, Marianne Rieckmann wrote: > Hi, > I have installed SystemImager as instructed by the manual (systemimager-manual-4.1.6.pdf) > using the RPM Based Distributions from the install script. With an image server on one machine > and a golden client on another. This is on a IBM blade cluster using CentOS6.2, with all the > DHCP, Domain name, tftp services working. > > When using the si_prepareclient, all looks good until the final part; > <pre> > This client is ready to have its image retrieved. You must now run > the "si_getimage" command on your imageserver. > > Your client has been successfully prepared. Boot kernel (copied from > this Linux distribution) and an initrd.img (generated by the > initrd_template package) can be found in /etc/systemimager/boot. > > Automatically create configuration file for systemconfigurator: > >> /etc/systemconfig/systemconfig.conf > FATAL: unable to detect an initrd for "2.6.32-220.17.1.el6.x86_64"! > WARNING: coulnd't create file /etc/systemconfig/systemconfig.conf (needed by systemconfigurator) > </pre> > > I think it is looking for one of the files in the /boot directory? > <pre> > [root@gouda6 boot]# ls > config-2.6.32-220.17.1.el6.x86_64 symvers-2.6.32-220.17.1.el6.x86_64.gz > config-2.6.32-220.el6.x86_64 symvers-2.6.32-220.el6.x86_64.gz > efi System.map-2.6.32-220.17.1.el6.x86_64 > grub System.map-2.6.32-220.el6.x86_64 > initramfs-2.6.32-220.17.1.el6.x86_64.img vmlinuz-2.6.32-220.17.1.el6.x86_64 > initramfs-2.6.32-220.el6.x86_64.img vmlinuz-2.6.32-220.el6.x86_64 > </pre> > > What do I need to do to get this to work? > Regards, > Marianne. > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > sisuite-users mailing list > sis...@li... > https://lists.sourceforge.net/lists/listinfo/sisuite-users > |
From: LAHAYE O. <oli...@ce...> - 2012-07-26 15:15:34
|
Thanks for the info, unfortunately, my problem was related to systemconfigurator (systemimager is now working fine with CentOS-6.2 and default kernel) I'll soon release an updated src.rpm of systemimager As you said, I used post install scripts to replace systemconfigurator and so far it works well. I just thaught that some one had fixed systemconfigurator or at least found a replacement so I could update systemimager master script. thanks for the input anyway. Olivier. ________________________________________ De : Bas van der Vlies [ba...@sa...] Date d'envoi : jeudi 26 juillet 2012 15:44 À : sis...@li... Objet : Re: [sisuite-users] RE : RE : Building SystemImager on CentOS 6.2? HELP > You can replace the SI kernel/initrd with the SALI kernel/initrd when > booting SI clients, and even run your old SI .master installation > scripts! Some minor tweaks to the .master script must be performed as > documented in > https://subtrac.sara.nl/oss/sali/wiki/SaliUsage/UsingYourOldInstallScript. > It really works! Of course, the installed CentOS 6.x client still > won't boot due to systemconfigurator :-( > > SALI has their own master scripts which are a lot simpler than the SI > scripts, but I think SALI hasn't been ported to CentOS 6.x yet because > 1) SALI still relies upon systemconfigurator, and 2) SALI doesn't seem > to support LVM. > Just some comments on SALI. We use the SALI kernel/initrd on Debian, Centos and SUSE Linux PowerPC. The development is mostly donein a Debian enviroment, Only when you want to make you own kernel/initrd you have to compile it from scratch. We do not use the systemimager-server tools anymore. They are replaced by the SALI server tools. they are noy documented yet :-( As for the systemconfigurator tools we have replaced them partially by functions in the `initrd` and postinst scripts. So we get rid of the systemconfigurator tools. We have some ideas howto replace them with other tools. As for LVM the SALI kernel/initrd support it. For now there are now functions for it to hide the complexity. So you must do the basic LVM commands in the master script. > In conclusion, it seems to me that if you can make systemconfigurator > work correctly on CentOS 6.x, then SI and/or SALI should work as well > for installing clients that will be bootable. > We have implemented a `grub2_install` functions or create a postinst script that will do it for you -- ******************************************************************** * Bas van der Vlies e-mail: ba...@sa... * * SARA - Academic Computing Services Amsterdam, The Netherlands * ******************************************************************** |
From: Bas v. d. V. <ba...@sa...> - 2012-07-26 13:57:24
|
> You can replace the SI kernel/initrd with the SALI kernel/initrd when > booting SI clients, and even run your old SI .master installation > scripts! Some minor tweaks to the .master script must be performed as > documented in > https://subtrac.sara.nl/oss/sali/wiki/SaliUsage/UsingYourOldInstallScript. > It really works! Of course, the installed CentOS 6.x client still > won't boot due to systemconfigurator :-( > > SALI has their own master scripts which are a lot simpler than the SI > scripts, but I think SALI hasn't been ported to CentOS 6.x yet because > 1) SALI still relies upon systemconfigurator, and 2) SALI doesn't seem > to support LVM. > Just some comments on SALI. We use the SALI kernel/initrd on Debian, Centos and SUSE Linux PowerPC. The development is mostly donein a Debian enviroment, Only when you want to make you own kernel/initrd you have to compile it from scratch. We do not use the systemimager-server tools anymore. They are replaced by the SALI server tools. they are noy documented yet :-( As for the systemconfigurator tools we have replaced them partially by functions in the `initrd` and postinst scripts. So we get rid of the systemconfigurator tools. We have some ideas howto replace them with other tools. As for LVM the SALI kernel/initrd support it. For now there are now functions for it to hide the complexity. So you must do the basic LVM commands in the master script. > In conclusion, it seems to me that if you can make systemconfigurator > work correctly on CentOS 6.x, then SI and/or SALI should work as well > for installing clients that will be bootable. > We have implemented a `grub2_install` functions or create a postinst script that will do it for you -- ******************************************************************** * Bas van der Vlies e-mail: ba...@sa... * * SARA - Academic Computing Services Amsterdam, The Netherlands * ******************************************************************** |
From: Ole H. N. <Ole...@fy...> - 2012-07-26 12:42:33
|
LAHAYE Olivier <oli...@ce...> wrote: > After trying to use my RPM on a CentOS-6.2 system, I'm facing some issues that are hard to fix, thus I'm asking before starting to try to make a update so it works on CentOS-6.2 > > 1st of all, I've already made some fix to systemimager so it works on CentOS-6.2, unfortunately, I'm stuck with systemconfigurator which is not supported on CentOS-6.2 (kernel-2.6) and therfore disable with no equivalent in systemimager. Because of this, the bootloader is not written to disk after the imaging, and therfore, the imaged system is unable to boot. > > So the actual situation is the following: > - systemimager-4.1.99.svn4568/lib/SystemImager/Server.pm:1788 SystemConfigurator deprecated and nothing to do its job (configure and write bootloader, ...) > > - systemconfigurator-2.2.11-1/lib/SystemConfig/{Initrd/RH.pm, Boot*, ???} need to be fixed to have kernel 2.6+ properly handeled (initramfs instead of initrd, dracut instead of initrd (compatible script exist for the moment), ... > > So now, I need help on having the ##SYSTEMCONFIGURATOR_PRE## and #SYSTEMCONFIGURATOR_POST## steps working in SystemImager. Works need to be done on the autoinstallscript and in systemconfigurator (or equivalent). I share your frustrations about systemconfigurator (and thus also SystemImager) not supporting CentOS 6.x! My experiences so far agree with yours: It's possible to tweak SystemImager to install a CentOS 6.x image onto a SI client's hard disk, but that image is unfortunately not bootable :-( Unfortunately I'm unable to help you with systemconfigurator. But I can add some observations regarding the use of SystemImager with CentOS 6.x (we're at 6.3 now) which reflect my experimentation to date: 1. SystemImager trunk as well as older versions CANNOT build on a CentOS 6.x system because several software packages (kernel, parted) are incompatible. The SI trunk won't even build on CentOS 5.x owing to the gzip package :-( 2. SystemImager packages built on a CentOS 5.x system CAN be installed correctly on CentOS 6.x SI server and golden client systems! Older versions of SI from CentOS 5.x seem to be working just fine on CentOS 6.x, and you can use si_getimage successfully. 3. A SI CentOS 6.x SI client machine can be installed correctly from the 6.x SI server in the usual way. However, at the end of the cloning process systemconfigurator gives some error messages, and the SI client is unable to boot :-( 4. If the SI clients have new hardware (Intel Sandy Bridge, for example), the SI kernel may be too old to work properly. Related to SI, the SALI project (https://subtrac.sara.nl/oss/sali) at SARA, Amsterdam offers a much newer kernel and initrd supporting modern filesystems and new hardware(?). You can replace the SI kernel/initrd with the SALI kernel/initrd when booting SI clients, and even run your old SI .master installation scripts! Some minor tweaks to the .master script must be performed as documented in https://subtrac.sara.nl/oss/sali/wiki/SaliUsage/UsingYourOldInstallScript. It really works! Of course, the installed CentOS 6.x client still won't boot due to systemconfigurator :-( SALI has their own master scripts which are a lot simpler than the SI scripts, but I think SALI hasn't been ported to CentOS 6.x yet because 1) SALI still relies upon systemconfigurator, and 2) SALI doesn't seem to support LVM. In conclusion, it seems to me that if you can make systemconfigurator work correctly on CentOS 6.x, then SI and/or SALI should work as well for installing clients that will be bootable. A totally different cloning approach would be to use RedHat's Kickstart - some of my colleagues have turned from SI to Kickstart already. If you add Cobbler (http://cobbler.github.com/) on top of Kickstart you may have quite a powerful cloning setup. Regards, Ole -- Ole Holm Nielsen Department of Physics, Technical University of Denmark |
From: LAHAYE O. <oli...@ce...> - 2012-07-25 12:50:53
|
Hi, After trying to use my RPM on a CentOS-6.2 system, I'm facing some issues that are hard to fix, thus I'm asking before starting to try to make a update so it works on CentOS-6.2 1st of all, I've already made some fix to systemimager so it works on CentOS-6.2, unfortunately, I'm stuck with systemconfigurator which is not supported on CentOS-6.2 (kernel-2.6) and therfore disable with no equivalent in systemimager. Because of this, the bootloader is not written to disk after the imaging, and therfore, the imaged system is unable to boot. So the actual situation is the following: - systemimager-4.1.99.svn4568/lib/SystemImager/Server.pm:1788 SystemConfigurator deprecated and nothing to do its job (configure and write bootloader, ...) - systemconfigurator-2.2.11-1/lib/SystemConfig/{Initrd/RH.pm, Boot*, ???} need to be fixed to have kernel 2.6+ properly handeled (initramfs instead of initrd, dracut instead of initrd (compatible script exist for the moment), ... So now, I need help on having the ##SYSTEMCONFIGURATOR_PRE## and #SYSTEMCONFIGURATOR_POST## steps working in SystemImager. Works need to be done on the autoinstallscript and in systemconfigurator (or equivalent). For now here are my patches for the latest systeminstaller to be able to deploy an image: - systemimager_initrd_coreutils_rul.patch Small Makefile fix (more clean) - systemimager_initrd_mklib_bef.patch Avoid overwriting libs that are alreadin installed in build_dir/ This prevent libmount-2.10 from CentOS-6.2 to overwrite libmount 2.12 from util-linux, and thus have mount, blkid and such to run instead of failing with wrong lib version. - systemimager_initrd_rul.patch Fix destination lib path. - systemimager_initrd_udev_rul.patch Disable systemd, libdir=/lib and use $MAKE - systemimager_initrd_util_linux_rul.patch Use $MAKE install instead of manualy copy binaries. If make install is not used, the libtool wrapper is copied instead of the real lib. - systemimager_server_pm.patch Fix that may be temporary. Right now, if the 1st partition inside the extended partition start at the same block, parted returns an error code. This fix avoid this by adding one block to the partition created inside the extended partition. Far from being a good fix, it makes at least things work. (fix commented in the patch) - systemimager_sysvinit_initrd_rul.patch Added a test to avoid failing when trying to copy /lib/ld* and no files are found. - systemimager_util_linux_rul.patch Enhanced configure step. Thanks a lot for your help. PS: Cross posted on oscar-devel mailing list ________________________________________ De : LAHAYE Olivier Date d'envoi : lundi 18 juin 2012 09:56 À : sis...@li... Objet : [sisuite-users] RE : Building SystemImager on CentOS 6.2? This is wired, as I'me sure I've faced the not updated kernel config issue (especially the compression method) and using my srpm on my centos-6.2, Id' doesn't ask for this. I've re-uploaded my src.rpm. please check with this rpm and all the sources. CentOS-6.2 and RHEL-6.2 should be very similar and shoud build the same way. (more over, this issue is related to the rpm, not the system. question is asked about things related to the rpm...) http://olivier.lahaye1.free.fr/SOURCES/ grab the external sources here an copy this in /usr/src Make sure you're using those sources when building and not something else downloaded elsewhere that could supersede some of those tarballs. http://olivier.lahaye1.free.fr/SRPMS/systemimager-4.2.0-0.6svn4568.el6.src.rpm Install this and build it as root (I know it is bad, but I hadn't time to fix the mknod issue). Cheers, Olivier. ________________________________________ De : Ole Holm Nielsen [Ole...@fy...] Date d'envoi : vendredi 15 juin 2012 22:25 À : sis...@li... Objet : Re: [sisuite-users] Building SystemImager on CentOS 6.2? LAHAYE Olivier <oli...@ce...> wrote: > Go there and fetch those sources. It should build without asking any questions (it does for me). > Hopefully I didn't forgot any Requires: in the spec file. > http://olivier.lahaye1.free.fr/SOURCES/ Thanks a lot! I grabbed the missing linux_2.6.34.1.orig.tar.gz file now. Unfortunately the build process still asks unwarranted questions. This is what I did: 1. I have checked out SVN trunk revision 4568 from https://systemimager.svn.sourceforge.net/svnroot/systemimager/trunk 2. In the trunk/ directory I do: make source_tarball 3. Then I build it: cd tmp; rpmbuild -ta systemimager-4.3.0.tar.bz2 However, during the build process the kernel wants to be reconfigured: > + make all > rm -rf /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1 > Extracting kernel source...done > cp -a /root/rpmbuild/BUILD/systemimager-4.3.0/patches/linux.x86_64.config /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1/.config > perl -pi -e "s/^EXTRAVERSION =\s*(\S*)\s*\n/EXTRAVERSION = \1-boel_v4.3.0\n/" /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1/Makefile > make -C /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1 oldconfig > make[1]: Entering directory `/root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1' > HOSTCC scripts/basic/fixdep > HOSTCC scripts/basic/docproc > HOSTCC scripts/basic/hash > HOSTCC scripts/kconfig/conf.o > HOSTCC scripts/kconfig/kxgettext.o > SHIPPED scripts/kconfig/zconf.tab.c > SHIPPED scripts/kconfig/lex.zconf.c > SHIPPED scripts/kconfig/zconf.hash.c > HOSTCC scripts/kconfig/zconf.tab.o > HOSTLD scripts/kconfig/conf > scripts/kconfig/conf -o arch/x86/Kconfig > * > * Restart config... > * > * > * General setup > * > Prompt for development and/or incomplete code/drivers (EXPERIMENTAL) [Y/n/?] y > Local version - append to kernel release (LOCALVERSION) [] > Automatically append version information to the version string (LOCALVERSION_AUTO) [Y/n/?] y > Kernel compression mode >> 1. Gzip (KERNEL_GZIP) > 2. Bzip2 (KERNEL_BZIP2) > 3. LZMA (KERNEL_LZMA) > 4. LZO (KERNEL_LZO) (NEW) > choice[1-4?]: So something must be wrong. I don't why I have to reconfigure the kernel whereas Olivier doesn't. FYI, my build host is RHEL6.2 (fully updated). Thanks for any help, Ole ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ sisuite-users mailing list sis...@li... https://lists.sourceforge.net/lists/listinfo/sisuite-users ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ sisuite-users mailing list sis...@li... https://lists.sourceforge.net/lists/listinfo/sisuite-users |
From: Marianne R. <mar...@ad...> - 2012-07-18 06:13:36
|
Hi, I have installed SystemImager as instructed by the manual (systemimager-manual-4.1.6.pdf) using the RPM Based Distributions from the install script. With an image server on one machine and a golden client on another. This is on a IBM blade cluster using CentOS6.2, with all the DHCP, Domain name, tftp services working. When using the si_prepareclient, all looks good until the final part; <pre> This client is ready to have its image retrieved. You must now run the "si_getimage" command on your imageserver. Your client has been successfully prepared. Boot kernel (copied from this Linux distribution) and an initrd.img (generated by the initrd_template package) can be found in /etc/systemimager/boot. Automatically create configuration file for systemconfigurator: >> /etc/systemconfig/systemconfig.conf FATAL: unable to detect an initrd for "2.6.32-220.17.1.el6.x86_64"! WARNING: coulnd't create file /etc/systemconfig/systemconfig.conf (needed by systemconfigurator) </pre> I think it is looking for one of the files in the /boot directory? <pre> [root@gouda6 boot]# ls config-2.6.32-220.17.1.el6.x86_64 symvers-2.6.32-220.17.1.el6.x86_64.gz config-2.6.32-220.el6.x86_64 symvers-2.6.32-220.el6.x86_64.gz efi System.map-2.6.32-220.17.1.el6.x86_64 grub System.map-2.6.32-220.el6.x86_64 initramfs-2.6.32-220.17.1.el6.x86_64.img vmlinuz-2.6.32-220.17.1.el6.x86_64 initramfs-2.6.32-220.el6.x86_64.img vmlinuz-2.6.32-220.el6.x86_64 </pre> What do I need to do to get this to work? Regards, Marianne. |
From: Damian M. <dam...@gm...> - 2012-07-10 17:50:17
|
Jan Groenewald <jan@...> writes:> > > > Hello All > > > Just ran si_prepareclient and the following error was reported. > > > ERROR unsupported kernel-3.0.0.0-12-generic > > > Is there an easy solution to this? > > > Thanks > > > David > > --- UseYourOwnKernel.pm.orig 2012-05-30 17:05:54.899344781 +0200 > +++ UseYourOwnKernel.pm 2012-05-31 09:06:09.331846666 +0200 Thanks for this patch! It works perfect for debian wheezy too. Is there a way to commit this patch to the svn or upgrade the SI? Thanks. |
From: Guy P. <gp...@ul...> - 2012-06-28 07:58:33
|
Hello Bjarne: >I met this issue recently. Probably, you have /boot on a separate >partition AND /boot on a lower/earlier number than / (right?). Yes: /boot on /dev/sda1 and / on /dev/sda2. >This is quite normal setup, but the present version(s) of SI-suite seems to >have problem with it. >What happens during installation (if I traced down and remember >correctly) is that the disk is that each partition is created, formatted >(with file system) and mounted on temporary mount point in turn. Thus, >assume that sda1 is /boot and sda2 is /, as in my own first try. >What happens is: > 1. sda1 is created and mounted as /a/boot/. > 2. sda2 is created and mounted as /a/ > >Note that pnt 2 obscures the mount point of sda1 (the boot partition). >Thus, during subsequent rsync, all files for /boot/ ends in /a/boot/, >which is really a directory on sda2(!) Thanks for your explanation. >My (ugly) solution is simply to get rid of the separate boot partition, >and let sda1 be /, which contains /boot as a directory. I think most >modern distributions should be happy with that choice. > >Let me know, how it works out. I actually went another way (perhaps uglier than yours): I downloaded a tarball from http://systemimager.svn.sourceforge.net/viewvc/systemimager/trunk/ Then I replaced /sbin/si_* scripts and *.pm in /usr/lib/systemimager/perl/SystemImager (as well as a few others) by those more recent than March 2008 (date of the "unstable" 4.1.6 rpm files) from the tarball. With this manipulation, the boot/ directory in the image tree was correctly populated by si_getimage. I tried to recompile the 4.1.6 sources after substituting all the tarball files more recent than 2008 but it crashed. Besides, I'm going to use the SALI kernel/initrd pair so I'm not utterly concerned by that unsuccessful recompilation. Regards, Guy Paulus >On 26-06-2012 3:45, Guy Paulus wrote: >> Hello: >> >> Let me outline the configurations of the golden client and the SI (and PXE) >> server: >> >> Golden client >> ------------- >> OS: Oracle Linux 6.2 >> Kernel version: 2.6.32-300.3.1.el6uek.x86_64 >> Disk layout: /dev/sda1 /boot ext4 >> /dev/sda2 / ext4 >> Version of systemimager: 4.1.6 >> Version of systemconfigurator: 2.2.11-1 >> >> SI server >> --------- >> OS: Scientific Linux 6.2 >> Kernel version: 2.6.32-220.7.1.el6.x86_64 >> Disk layout: /boot (ext4) + LVM [LVM version: 2.02.87(2)-RHEL6 (2011-10-12); >> Library version: 1.02.66-RHEL6 (2011-10-12); Driver version: 4.22.6] >> Version of systemimager: 4.1.6 >> Version of systemconfigurator: 2.2.11-1 >> Using rsync to pull the golden client image >> >> Problem >> ------- >> After performing si_getimage on the SI server, the <image>/boot directory is >> empty. As a result, when I PXE boot a client, the autoinstall fails because >> systemconfigurator complains that the kernel >> /boot/vmlinuz-2.6.32-300.3.1.el6uek.x86_64 does not exist. >> There is no explicit exclusion of /boot in /etc/systemimager/getimage.exclude. >> >> Did I miss something at the si_prepareclient stage ? I did >> si_prepareclient --server <ip address> --my-modules --yes >> >> The /etc/systemconfig/systemconfig.conf file reads >> >> [BOOT] >> ROOTDEV = /dev/sda2 >> BOOTDEV = /dev/sda >> DEFAULTBOOT = systemimager >> >> [KERNEL0] >> LABEL = systemimager >> PATH = /boot/vmlinuz-2.6.32-300.3.1.el6uek.x86_64 >> INITRD = /boot/initrd-2.6.32-300.3.1.el6uek.x86_64.img >> APPEND = KEYBOARDTYPE=pc KEYTABLE=us LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 quiet rd_NO_DM rd_NO_LUKS rd_NO_LVM rd_NO_MD rhgb ro >> >> Or did I miss something at the si_getimage stage ? I did >> >> sudo si_getimage --golden-client <ip address> --image blade0 --ip-assignment replicant --post-install reboot >> >> Thanks. >> >> Guy Paulus >> >> >> ------------------------------------------------------------------------------ >> Live Security Virtual Conference >> Exclusive live event will cover all the ways today's security and >> threat landscape has changed and how IT managers can respond. Discussions >> will include endpoint security, mobile security and the latest in malware >> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ >> _______________________________________________ >> sisuite-users mailing list >> sis...@li... >> https://lists.sourceforge.net/lists/listinfo/sisuite-users >> > > > > |
From: Büchmann H. <en...@bu...> - 2012-06-28 07:52:14
|
Hi, For various reasons, I want to use short host names for my nodes. This means that /etc/sysconfig/network on a machine (after rolling out an image) should have a short host name (e.g. HOSTNAME=n001) rather than a fully qualified name (e.g. HOSTNAME=dn001.mydomain.net). SystemConfigurator supports this by adding "SHORTHOSTNAME = YES" to the [NETWORK] section in /etc/systemconfig/systemconfig.conf However, AFAIK si_prepareclient does not (under any circumstances) even write the [NETWORK] section as it creates a new systemconfig.conf (overwriting the old, obviously). Which means that I probably cannot configure my way out of the issue using si-suite commands. I this correctly understood? An obvious solution is to edit /etc/systemconfig/systemconfig.conf on my golden node after running si_prepareclient. It is simple, as I just append a few well-known lines to the file. But maybe it would be even better, if si_prepareclient was able to do it? (I do not have the full understanding of how si_prepareclient etc works, so I may not be the right person to do it. But maybe adding a few more command-line options is not too difficult? All input is welcome. Thanks, Bjarne |
From: Büchmann H. <en...@bu...> - 2012-06-28 07:09:04
|
Hi Guy, I met this issue recently. Probably, you have /boot on a separate partition AND /boot on a lower/earlier number than / (right?). This is quite normal setup, but the present version(s) of SI-suite seems to have problem with it. What happens during installation (if I traced down and remember correctly) is that the disk is that each partition is created, formatted (with file system) and mounted on temporary mount point in turn. Thus, assume that sda1 is /boot and sda2 is /, as in my own first try. What happens is: 1. sda1 is created and mounted as /a/boot/. 2. sda2 is created and mounted as /a/ Note that pnt 2 obscures the mount point of sda1 (the boot partition). Thus, during subsequent rsync, all files for /boot/ ends in /a/boot/, which is really a directory on sda2(!) My (ugly) solution is simply to get rid of the separate boot partition, and let sda1 be /, which contains /boot as a directory. I think most modern distributions should be happy with that choice. Let me know, how it works out. /Bjarne On 26-06-2012 3:45, Guy Paulus wrote: > Hello: > > Let me outline the configurations of the golden client and the SI (and PXE) > server: > > Golden client > ------------- > OS: Oracle Linux 6.2 > Kernel version: 2.6.32-300.3.1.el6uek.x86_64 > Disk layout: /dev/sda1 /boot ext4 > /dev/sda2 / ext4 > Version of systemimager: 4.1.6 > Version of systemconfigurator: 2.2.11-1 > > SI server > --------- > OS: Scientific Linux 6.2 > Kernel version: 2.6.32-220.7.1.el6.x86_64 > Disk layout: /boot (ext4) + LVM [LVM version: 2.02.87(2)-RHEL6 (2011-10-12); > Library version: 1.02.66-RHEL6 (2011-10-12); Driver version: 4.22.6] > Version of systemimager: 4.1.6 > Version of systemconfigurator: 2.2.11-1 > Using rsync to pull the golden client image > > Problem > ------- > After performing si_getimage on the SI server, the <image>/boot directory is > empty. As a result, when I PXE boot a client, the autoinstall fails because > systemconfigurator complains that the kernel > /boot/vmlinuz-2.6.32-300.3.1.el6uek.x86_64 does not exist. > There is no explicit exclusion of /boot in /etc/systemimager/getimage.exclude. > > Did I miss something at the si_prepareclient stage ? I did > si_prepareclient --server <ip address> --my-modules --yes > > The /etc/systemconfig/systemconfig.conf file reads > > [BOOT] > ROOTDEV = /dev/sda2 > BOOTDEV = /dev/sda > DEFAULTBOOT = systemimager > > [KERNEL0] > LABEL = systemimager > PATH = /boot/vmlinuz-2.6.32-300.3.1.el6uek.x86_64 > INITRD = /boot/initrd-2.6.32-300.3.1.el6uek.x86_64.img > APPEND = KEYBOARDTYPE=pc KEYTABLE=us LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 quiet rd_NO_DM rd_NO_LUKS rd_NO_LVM rd_NO_MD rhgb ro > > Or did I miss something at the si_getimage stage ? I did > > sudo si_getimage --golden-client <ip address> --image blade0 --ip-assignment replicant --post-install reboot > > Thanks. > > Guy Paulus > > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > sisuite-users mailing list > sis...@li... > https://lists.sourceforge.net/lists/listinfo/sisuite-users > |
From: Büchmann H. <en...@bu...> - 2012-06-28 07:08:56
|
Hi, I think the short answer to this query is "don't use UUID for mounting in fstab with si-suite". You cannot control the UUID when the partition is formatted. Use simple device numbers (sda1 etc) or labels. In the latter case, you may have to tweak the installation script, or talk to the SALI guys about their extensions. Best, Bjarne On 20-06-2012 6:27, E-Blokos wrote: > Hi, > I update my kernel to 3.2.21 > but now when I do si_prepareclient --server nodexxx > I get > Using "parted" to gather information about disk: > /dev/sda > Using "parted" to gather information about disk: > /dev/sdb > > WARNING: unable to identify the device with > "UUID=92a2b62c-6cfc-440f-aa90-d922523c1e62" defined in > /etc/fstab! > Manually set the "real_dev" and "format" attributes in > "/etc/systemimager/autoinstallscript.conf" to properly create > this device > during the autoinstall process. > ERROR: unsupported kernel 3.2.21-rt32! > can I correct this manually ? > Thanks > Franck |
From: Guy P. <gp...@ul...> - 2012-06-26 14:20:27
|
Hello: Let me outline the configurations of the golden client and the SI (and PXE) server: Golden client ------------- OS: Oracle Linux 6.2 Kernel version: 2.6.32-300.3.1.el6uek.x86_64 Disk layout: /dev/sda1 /boot ext4 /dev/sda2 / ext4 Version of systemimager: 4.1.6 Version of systemconfigurator: 2.2.11-1 SI server --------- OS: Scientific Linux 6.2 Kernel version: 2.6.32-220.7.1.el6.x86_64 Disk layout: /boot (ext4) + LVM [LVM version: 2.02.87(2)-RHEL6 (2011-10-12); Library version: 1.02.66-RHEL6 (2011-10-12); Driver version: 4.22.6] Version of systemimager: 4.1.6 Version of systemconfigurator: 2.2.11-1 Using rsync to pull the golden client image Problem ------- After performing si_getimage on the SI server, the <image>/boot directory is empty. As a result, when I PXE boot a client, the autoinstall fails because systemconfigurator complains that the kernel /boot/vmlinuz-2.6.32-300.3.1.el6uek.x86_64 does not exist. There is no explicit exclusion of /boot in /etc/systemimager/getimage.exclude. Did I miss something at the si_prepareclient stage ? I did si_prepareclient --server <ip address> --my-modules --yes The /etc/systemconfig/systemconfig.conf file reads [BOOT] ROOTDEV = /dev/sda2 BOOTDEV = /dev/sda DEFAULTBOOT = systemimager [KERNEL0] LABEL = systemimager PATH = /boot/vmlinuz-2.6.32-300.3.1.el6uek.x86_64 INITRD = /boot/initrd-2.6.32-300.3.1.el6uek.x86_64.img APPEND = KEYBOARDTYPE=pc KEYTABLE=us LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 quiet rd_NO_DM rd_NO_LUKS rd_NO_LVM rd_NO_MD rhgb ro Or did I miss something at the si_getimage stage ? I did sudo si_getimage --golden-client <ip address> --image blade0 --ip-assignment replicant --post-install reboot Thanks. Guy Paulus |
From: Ole H. N. <Ole...@fy...> - 2012-06-21 12:50:05
|
Unfortunately no one has yet offered any advice for building SystemImager (trunk version) on CentOS 6.2. Out of desperation I tried to build the trunk version on the RHEL 5.8 server where we have used SystemImager 4.1.7 successfully for years. Again, the SI trunk version fails to build on RHEL 5.8 :-( The kernel configuration again asks a number of configuration questions, but I just choose the default values to proceed. Finally, the build process crashed at this stage: cd /usr/src/redhat/BUILD/systemimager-4.3.0/initrd_source/src && tar -xvJf gzip-1.4.tar.xz tar: invalid option -- J Try `tar --help' or `tar --usage' for more information. make: *** [/usr/src/redhat/BUILD/systemimager-4.3.0/initrd_source/src/gzip-1.4.unpack] Error 64 error: Bad exit status from /var/tmp/rpm-tmp.6420 (%build) The problem is the file gzip-1.4.tar.xz which can't be unpacked on RHEL5/CentOS5 because the tar command on such systems don't know the .xz type of compression (tar -J command). Could someone knowledgeable about the SI trunk kindly replace the gzip-1.4.tar.xz package by the similar gzip-1.4.tar.gz package and change the -J flag into -z ? Thanks a lot, Ole |
From: E-Blokos <in...@e-...> - 2012-06-20 17:02:58
|
Hi, I update my kernel to 3.2.21 but now when I do si_prepareclient --server nodexxx I get Using "parted" to gather information about disk: /dev/sda Using "parted" to gather information about disk: /dev/sdb WARNING: unable to identify the device with "UUID=92a2b62c-6cfc-440f-aa90-d922523c1e62" defined in /etc/fstab! Manually set the "real_dev" and "format" attributes in "/etc/systemimager/autoinstallscript.conf" to properly create this device during the autoinstall process. ERROR: unsupported kernel 3.2.21-rt32! can I correct this manually ? Thanks Franck |
From: Ole H. N. <Ole...@fy...> - 2012-06-18 10:07:07
|
More problems in building SystemImager on CentOS 6.2: Besides the kernel reconfiguration issue (which may be innocent, perhaps?), the building of parted 2.3 aborts on RHEL 6.2 during the configuration stage: In /root/rpmbuild/BUILD/systemimager-4.3.0/initrd_source/src/parted-2.3 the config.log shows: # ./configure --disable-Werror --disable-shared --prefix= ... checking for dm_task_create in -ldevmapper... no configure: error: libdevmapper could not be found, but is required for the --enable-device-mapper option, which is enabled by default. Either disable device-mapper support with --disable-device-mapper or download and install device-mapper from: http://sources.redhat.com/dm/ Note: if you are using precompiled packages you will need the development package as well (it may be called device-mapper-devel or something similar). The /lib/libdevmapper.so.1.02 shared library does exist. But RHEL6 does NOT have any package named device-mapper-devel or the like. The check dm_task_create in -ldevmapper can be extracted from the configure script: #ifdef __cplusplus extern "C" #endif char dm_task_create (); int main () { return dm_task_create (); ; return 0; } Compiling this on RHEL6 gives: # gcc test.c -ldevmapper /usr/bin/ld: cannot find -ldevmapper collect2: ld returned 1 exit status I'm really confused: Why does this code break, while other people seem to have built SystemImager (including parted) successfully on RHEL6/CentOS6 ? Thanks for sharing any insights, Ole |
From: LAHAYE O. <oli...@ce...> - 2012-06-18 07:56:28
|
This is wired, as I'me sure I've faced the not updated kernel config issue (especially the compression method) and using my srpm on my centos-6.2, Id' doesn't ask for this. I've re-uploaded my src.rpm. please check with this rpm and all the sources. CentOS-6.2 and RHEL-6.2 should be very similar and shoud build the same way. (more over, this issue is related to the rpm, not the system. question is asked about things related to the rpm...) http://olivier.lahaye1.free.fr/SOURCES/ grab the external sources here an copy this in /usr/src Make sure you're using those sources when building and not something else downloaded elsewhere that could supersede some of those tarballs. http://olivier.lahaye1.free.fr/SRPMS/systemimager-4.2.0-0.6svn4568.el6.src.rpm Install this and build it as root (I know it is bad, but I hadn't time to fix the mknod issue). Cheers, Olivier. ________________________________________ De : Ole Holm Nielsen [Ole...@fy...] Date d'envoi : vendredi 15 juin 2012 22:25 À : sis...@li... Objet : Re: [sisuite-users] Building SystemImager on CentOS 6.2? LAHAYE Olivier <oli...@ce...> wrote: > Go there and fetch those sources. It should build without asking any questions (it does for me). > Hopefully I didn't forgot any Requires: in the spec file. > http://olivier.lahaye1.free.fr/SOURCES/ Thanks a lot! I grabbed the missing linux_2.6.34.1.orig.tar.gz file now. Unfortunately the build process still asks unwarranted questions. This is what I did: 1. I have checked out SVN trunk revision 4568 from https://systemimager.svn.sourceforge.net/svnroot/systemimager/trunk 2. In the trunk/ directory I do: make source_tarball 3. Then I build it: cd tmp; rpmbuild -ta systemimager-4.3.0.tar.bz2 However, during the build process the kernel wants to be reconfigured: > + make all > rm -rf /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1 > Extracting kernel source...done > cp -a /root/rpmbuild/BUILD/systemimager-4.3.0/patches/linux.x86_64.config /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1/.config > perl -pi -e "s/^EXTRAVERSION =\s*(\S*)\s*\n/EXTRAVERSION = \1-boel_v4.3.0\n/" /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1/Makefile > make -C /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1 oldconfig > make[1]: Entering directory `/root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1' > HOSTCC scripts/basic/fixdep > HOSTCC scripts/basic/docproc > HOSTCC scripts/basic/hash > HOSTCC scripts/kconfig/conf.o > HOSTCC scripts/kconfig/kxgettext.o > SHIPPED scripts/kconfig/zconf.tab.c > SHIPPED scripts/kconfig/lex.zconf.c > SHIPPED scripts/kconfig/zconf.hash.c > HOSTCC scripts/kconfig/zconf.tab.o > HOSTLD scripts/kconfig/conf > scripts/kconfig/conf -o arch/x86/Kconfig > * > * Restart config... > * > * > * General setup > * > Prompt for development and/or incomplete code/drivers (EXPERIMENTAL) [Y/n/?] y > Local version - append to kernel release (LOCALVERSION) [] > Automatically append version information to the version string (LOCALVERSION_AUTO) [Y/n/?] y > Kernel compression mode >> 1. Gzip (KERNEL_GZIP) > 2. Bzip2 (KERNEL_BZIP2) > 3. LZMA (KERNEL_LZMA) > 4. LZO (KERNEL_LZO) (NEW) > choice[1-4?]: So something must be wrong. I don't why I have to reconfigure the kernel whereas Olivier doesn't. FYI, my build host is RHEL6.2 (fully updated). Thanks for any help, Ole ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ sisuite-users mailing list sis...@li... https://lists.sourceforge.net/lists/listinfo/sisuite-users |
From: Andrej <and...@gm...> - 2012-06-17 21:41:16
|
On 16 June 2012 07:37, Damian Montaldo <dam...@gm...> wrote: > Yes and it was a headache because it's very old and deprecated to > install in a up-to-date squeeze. > I finally install it in squeeze but I've ended with SALI. > > If you really really want to use it I can give you some help. Thanks for your kind offer! :) > Try with these deb in the debian archive > wget http://archive.debian.org/debian/pool/main/c/cdrkit/mkisofs_1.1.9-1_all.deb > > the run as root or with sudo > dpkg --install mkisofs_1.1.9-1_all.deb Cool - those steps took me past the failure on the dependency check. \> Then you have to install genisoimage because the last versión of > mkisofs it's only a wrapper to geniso (the transitional package) Already had that installed in the hopes that it might register under the "old name" as well. With another extra package thrown in (apt-get install makedev; since udev does all the lifting these days it's not installed by default) I seem to have successfully installed the packages I think I need :) Will now try to configure it. > Damian. Cheers, Andrej |
From: Ole H. N. <Ole...@fy...> - 2012-06-15 20:25:45
|
LAHAYE Olivier <oli...@ce...> wrote: > Go there and fetch those sources. It should build without asking any questions (it does for me). > Hopefully I didn't forgot any Requires: in the spec file. > http://olivier.lahaye1.free.fr/SOURCES/ Thanks a lot! I grabbed the missing linux_2.6.34.1.orig.tar.gz file now. Unfortunately the build process still asks unwarranted questions. This is what I did: 1. I have checked out SVN trunk revision 4568 from https://systemimager.svn.sourceforge.net/svnroot/systemimager/trunk 2. In the trunk/ directory I do: make source_tarball 3. Then I build it: cd tmp; rpmbuild -ta systemimager-4.3.0.tar.bz2 However, during the build process the kernel wants to be reconfigured: > + make all > rm -rf /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1 > Extracting kernel source...done > cp -a /root/rpmbuild/BUILD/systemimager-4.3.0/patches/linux.x86_64.config /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1/.config > perl -pi -e "s/^EXTRAVERSION =\s*(\S*)\s*\n/EXTRAVERSION = \1-boel_v4.3.0\n/" /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1/Makefile > make -C /root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1 oldconfig > make[1]: Entering directory `/root/rpmbuild/BUILD/systemimager-4.3.0/src/linux-2.6.34.1' > HOSTCC scripts/basic/fixdep > HOSTCC scripts/basic/docproc > HOSTCC scripts/basic/hash > HOSTCC scripts/kconfig/conf.o > HOSTCC scripts/kconfig/kxgettext.o > SHIPPED scripts/kconfig/zconf.tab.c > SHIPPED scripts/kconfig/lex.zconf.c > SHIPPED scripts/kconfig/zconf.hash.c > HOSTCC scripts/kconfig/zconf.tab.o > HOSTLD scripts/kconfig/conf > scripts/kconfig/conf -o arch/x86/Kconfig > * > * Restart config... > * > * > * General setup > * > Prompt for development and/or incomplete code/drivers (EXPERIMENTAL) [Y/n/?] y > Local version - append to kernel release (LOCALVERSION) [] > Automatically append version information to the version string (LOCALVERSION_AUTO) [Y/n/?] y > Kernel compression mode >> 1. Gzip (KERNEL_GZIP) > 2. Bzip2 (KERNEL_BZIP2) > 3. LZMA (KERNEL_LZMA) > 4. LZO (KERNEL_LZO) (NEW) > choice[1-4?]: So something must be wrong. I don't why I have to reconfigure the kernel whereas Olivier doesn't. FYI, my build host is RHEL6.2 (fully updated). Thanks for any help, Ole |
From: Damian M. <dam...@gm...> - 2012-06-15 19:37:57
|
On Thu, Jun 14, 2012 at 8:54 PM, Andrej <and...@gm...> wrote: > Hi, > > I have been trying to get this combo (see subject) to play nice, but > can't seem to get anywhere ... > > Has anyone managed to install SystemImager on squeeze?\ Yes and it was a headache because it's very old and deprecated to install in a up-to-date squeeze. I finally install it in squeeze but I've ended with SALI. If you really really want to use it I can give you some help. > The packages from SourceForge fail because of unmet dependencies > (namely genisofs having replaced mkisofs, and wodim cdreord are > the culprits) .... Try with these deb in the debian archive wget http://archive.debian.org/debian/pool/main/c/cdrkit/mkisofs_1.1.9-1_all.deb the run as root or with sudo dpkg --install mkisofs_1.1.9-1_all.deb Then you have to install genisoimage because the last versión of mkisofs it's only a wrapper to geniso (the transitional package) > Compilation from source fails of a supposedly missing libuuid, which, in > fact, is installed (64 & 32 bit), along w/ e2progs-dev for headers > and such ... I still have the deb from sarge I think or may be from before but if you grab the debs from the archive I think it should works too. Damian. |
From: Brian F. <br...@th...> - 2012-06-15 14:21:09
|
We want to eliminate the need for mknod, definitely. I would still consider SVN trunk to be in a testing state. Most of the intelligence is the same, but the way the autoinstall client binaries and kernel are built is a bit different. Please let me know how it goes. -Brian On 06/14/2012 09:30 AM, LAHAYE Olivier wrote: > > I've built the rpm on CentOS-6.2. The resulting rpms seems ok. As I'm porting OSCAR on CentOS-6.2, I've been able to build client images so far, but I can't tell if the resulting image is fully functionnal as systemimage-oscar fail to generate the KERNEL0 line in /var/lib/systemimager/images/<imagename>/etc/systemconfig/systemconfig.conf (problem outside these rpms) > > To build the rpms, just look for the missing archive names in google and download the archive into /usr./src directory. > Note that (sadly) you have to be root to build the rpms as the rpm needs mknode if I'm correct and I didn't find time to look at this issue and have the rpm buildable as non-root. > > Although, it would be cool if the source repository could be updated as it's a pain in the ass to google and fetch the correct sources archives and copy those in the /usr/src directory. > > Anybody able to do that? > > Regards, > -- Brian Elliott Finley Mobile: 469.444.0167 |
From: LAHAYE O. <oli...@ce...> - 2012-06-15 13:23:53
|
Go there and fetch those sources. It should build without asking any questions (it does for me). Hopefully I didn't forgot any Requires: in the spec file. http://olivier.lahaye1.free.fr/SOURCES/ Olivier. ________________________________________ De : Ole Holm Nielsen [Ole...@fy...] Date d'envoi : vendredi 15 juin 2012 15:05 À : sis...@li... Objet : Re: [sisuite-users] Building SystemImager on CentOS 6.2? LAHAYE Olivier <oli...@ce...> wrote: > I've built the rpm on CentOS-6.2. The resulting rpms seems ok. As I'm porting OSCAR on CentOS-6.2, I've been able to build client images so far, but I can't tell if the resulting image is fully functionnal as systemimage-oscar fail to generate the KERNEL0 line in /var/lib/systemimager/images/<imagename>/etc/systemconfig/systemconfig.conf (problem outside these rpms) > > To build the rpms, just look for the missing archive names in google and download the archive into /usr./src directory. > Note that (sadly) you have to be root to build the rpms as the rpm needs mknode if I'm correct and I didn't find time to look at this issue and have the rpm buildable as non-root. > > Although, it would be cool if the source repository could be updated as it's a pain in the ass to google and fetch the correct sources archives and copy those in the /usr/src directory. I've made a bit of progress towards building the SI RPMs on an RHEL6.2 server, but I didn't make it to the goal yet: 1. I copied the downloaded sources from /root/rpmbuild/BUILD/systemimager-4.3.0/initrd_source/src/ to the permanent place /usr/src/. Strangely, today the correct e2fsprogs can be downloaded successfully (it was absent yesterday). 2. The next missing download is a kernel source file http://download.systemimager.org/pub/linux/linux_2.6.34.1.orig.tar.gz. So I grabbed a similar file from http://www.kernel.org/pub/linux/kernel/v2.6/ and renamed it with the .orig (probably a no-no!). The kernel.org version seems to build, but I had to answer a bunch of configure questions. 3. Next failure occurs when building parted-2.3: ... checking for dm_task_create in -ldevmapper... no configure: error: libdevmapper could not be found, but is required for the --enable-device-mapper option, which is enabled by default. Either disable device-mapper support with --disable-device-mapper or download and install device-mapper from: http://sources.redhat.com/dm/ FYI, our RHEL6.2 server does have the device-mapper 32- and 64-bit RPMs installed with /lib/libdevmapper.so.1.02. Question: Am I on the right track, or is there a better way to build the SI RPMs? Thanks a lot, Ole -- Ole Holm Nielsen Department of Physics, Technical University of Denmark ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ sisuite-users mailing list sis...@li... https://lists.sourceforge.net/lists/listinfo/sisuite-users |
From: Ole H. N. <Ole...@fy...> - 2012-06-15 13:06:06
|
LAHAYE Olivier <oli...@ce...> wrote: > I've built the rpm on CentOS-6.2. The resulting rpms seems ok. As I'm porting OSCAR on CentOS-6.2, I've been able to build client images so far, but I can't tell if the resulting image is fully functionnal as systemimage-oscar fail to generate the KERNEL0 line in /var/lib/systemimager/images/<imagename>/etc/systemconfig/systemconfig.conf (problem outside these rpms) > > To build the rpms, just look for the missing archive names in google and download the archive into /usr./src directory. > Note that (sadly) you have to be root to build the rpms as the rpm needs mknode if I'm correct and I didn't find time to look at this issue and have the rpm buildable as non-root. > > Although, it would be cool if the source repository could be updated as it's a pain in the ass to google and fetch the correct sources archives and copy those in the /usr/src directory. I've made a bit of progress towards building the SI RPMs on an RHEL6.2 server, but I didn't make it to the goal yet: 1. I copied the downloaded sources from /root/rpmbuild/BUILD/systemimager-4.3.0/initrd_source/src/ to the permanent place /usr/src/. Strangely, today the correct e2fsprogs can be downloaded successfully (it was absent yesterday). 2. The next missing download is a kernel source file http://download.systemimager.org/pub/linux/linux_2.6.34.1.orig.tar.gz. So I grabbed a similar file from http://www.kernel.org/pub/linux/kernel/v2.6/ and renamed it with the .orig (probably a no-no!). The kernel.org version seems to build, but I had to answer a bunch of configure questions. 3. Next failure occurs when building parted-2.3: ... checking for dm_task_create in -ldevmapper... no configure: error: libdevmapper could not be found, but is required for the --enable-device-mapper option, which is enabled by default. Either disable device-mapper support with --disable-device-mapper or download and install device-mapper from: http://sources.redhat.com/dm/ FYI, our RHEL6.2 server does have the device-mapper 32- and 64-bit RPMs installed with /lib/libdevmapper.so.1.02. Question: Am I on the right track, or is there a better way to build the SI RPMs? Thanks a lot, Ole -- Ole Holm Nielsen Department of Physics, Technical University of Denmark |
From: Andrej <and...@gm...> - 2012-06-14 23:54:50
|
Hi, I have been trying to get this combo (see subject) to play nice, but can't seem to get anywhere ... Has anyone managed to install SystemImager on squeeze? The packages from SourceForge fail because of unmet dependencies (namely genisofs having replaced mkisofs, and wodim cdreord are the culprits) .... Compilation from source fails of a supposedly missing libuuid, which, in fact, is installed (64 & 32 bit), along w/ e2progs-dev for headers and such ... Cheers, Andrej |
From: Bernard Li <be...@va...> - 2012-06-14 18:34:43
|
If you guys could attach the patch here I can take a stab at checking it in. Olivier, would you be interested in getting commit access to the SVN repo? Thanks, Bernard On Thu, Jun 14, 2012 at 8:30 AM, LAHAYE Olivier <oli...@ce...> wrote: > > I've built the rpm on CentOS-6.2. The resulting rpms seems ok. As I'm porting OSCAR on CentOS-6.2, I've been able to build client images so far, but I can't tell if the resulting image is fully functionnal as systemimage-oscar fail to generate the KERNEL0 line in /var/lib/systemimager/images/<imagename>/etc/systemconfig/systemconfig.conf (problem outside these rpms) > > To build the rpms, just look for the missing archive names in google and download the archive into /usr./src directory. > Note that (sadly) you have to be root to build the rpms as the rpm needs mknode if I'm correct and I didn't find time to look at this issue and have the rpm buildable as non-root. > > Although, it would be cool if the source repository could be updated as it's a pain in the ass to google and fetch the correct sources archives and copy those in the /usr/src directory. > > Anybody able to do that? > > Regards, > > -- > Olivier. > ________________________________________ > De : Ole Holm Nielsen [Ole...@fy...] > Date d'envoi : jeudi 14 juin 2012 15:28 > À : sis...@li... > Objet : [sisuite-users] Building SystemImager on CentOS 6.2? > > Hi SystemImager guys, > > We would like to build SystemImager on RedHat 6.2 / CentOS 6.2. I got > the latest SI rev. 4568 from SVN by: > svn checkout > https://systemimager.svn.sourceforge.net/svnroot/systemimager/trunk > (the website http://www.systemimager.org/ is still sadly incomplete). > > However, additional tweaks seem to be required for CentOS 6.2, see > http://sourceforge.net/mailarchive/forum.php?thread_name=01EDEEAE5DA8584CAAD628AA5B06179A048D471F%40EXDAG0-A3.intra.cea.fr&forum_name=sisuite-users > My RPM build crashes because of a download reference to an obsolete > e2fsprogs version. > > Questions: > * Does anyone have complete and detailed SI build instructions for > CentOS 6.2? > * Can anyone with write access to SVN commit the changes required for 6.2? > > Thanks, > Ole > > -- > Ole Holm Nielsen > Department of Physics, Technical University of Denmark > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > sisuite-users mailing list > sis...@li... > https://lists.sourceforge.net/lists/listinfo/sisuite-users > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > sisuite-users mailing list > sis...@li... > https://lists.sourceforge.net/lists/listinfo/sisuite-users |
From: LAHAYE O. <oli...@ce...> - 2012-06-14 15:30:43
|
I've built the rpm on CentOS-6.2. The resulting rpms seems ok. As I'm porting OSCAR on CentOS-6.2, I've been able to build client images so far, but I can't tell if the resulting image is fully functionnal as systemimage-oscar fail to generate the KERNEL0 line in /var/lib/systemimager/images/<imagename>/etc/systemconfig/systemconfig.conf (problem outside these rpms) To build the rpms, just look for the missing archive names in google and download the archive into /usr./src directory. Note that (sadly) you have to be root to build the rpms as the rpm needs mknode if I'm correct and I didn't find time to look at this issue and have the rpm buildable as non-root. Although, it would be cool if the source repository could be updated as it's a pain in the ass to google and fetch the correct sources archives and copy those in the /usr/src directory. Anybody able to do that? Regards, -- Olivier. ________________________________________ De : Ole Holm Nielsen [Ole...@fy...] Date d'envoi : jeudi 14 juin 2012 15:28 À : sis...@li... Objet : [sisuite-users] Building SystemImager on CentOS 6.2? Hi SystemImager guys, We would like to build SystemImager on RedHat 6.2 / CentOS 6.2. I got the latest SI rev. 4568 from SVN by: svn checkout https://systemimager.svn.sourceforge.net/svnroot/systemimager/trunk (the website http://www.systemimager.org/ is still sadly incomplete). However, additional tweaks seem to be required for CentOS 6.2, see http://sourceforge.net/mailarchive/forum.php?thread_name=01EDEEAE5DA8584CAAD628AA5B06179A048D471F%40EXDAG0-A3.intra.cea.fr&forum_name=sisuite-users My RPM build crashes because of a download reference to an obsolete e2fsprogs version. Questions: * Does anyone have complete and detailed SI build instructions for CentOS 6.2? * Can anyone with write access to SVN commit the changes required for 6.2? Thanks, Ole -- Ole Holm Nielsen Department of Physics, Technical University of Denmark ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ sisuite-users mailing list sis...@li... https://lists.sourceforge.net/lists/listinfo/sisuite-users |