So I found the reason it was giving me the "kernel too old" message: when running the line
chroot /a/ systemconfigurator --version from the master script, chroot fails with that message. It seems to be related to different glibc versions on the BOEL root and the chrooted environment. On the BOEL root, glibc exists in the /lib folder: /lib # ./libc.so.6 GNU C Library stable release version 2.12 (snipped for brevity) however, on /a/ it only exists in /a/lib64/. Trying to run /a/lib64/libc.so.6 (to check the version) results in the following message: /a/lib64 # ./libc.so.6 ./libc.so.6: relocation error: ./libc.so.6: symbol _dl_starting_up, version GLIBC_PRIVATE not defined in ld-linux-x86-64.so.2 with link time reference On my golden client, running ./libc.so.6 succeeds: [root@test bin]$ /lib64/libc.so.6 GNU C Library development release version 2.13.90, by Roland McGrath et al. So there is a version mismatch going on here, although that must be common, so I'm not sure how to proceed. Thanks, Tom Yue On 5/18/2011 2:20 PM, Tom Yue wrote: > Hi all, > > I'm trying to use SystemImager 4.1.99 (the SVN copy that Bernard > currently has in his folder) to deploy a Fedora 15 beta golden client to > a node. The following packages are installed on the golden client: > > perl-AppConfig-1.66-10.fc15.noarch > systemconfigurator-2.2.11-1.noarch > systemimager-common-4.1.99.svn4556_bli-1.noarch > systemimager-client-4.1.99.svn4556_bli-1.noarch > systemimager-x86_64initrd_template-4.1.99.svn4556_bli-1.noarch > > and the following packages are installed on the image server, running > Fedora 14: > > perl-AppConfig-1.66-8.fc14.noarch > systemconfigurator-2.2.11-1.noarch > systemimager-common-4.1.99.svn4556_bli-1.noarch > systemimager-server-4.1.99.svn4556_bli-1.noarch > systemimager-x86_64boot-standard-4.1.99.svn4556_bli-1.noarch > > The image server, the golden client and the node are all running the > same spec, which is a fairly modern SuperMicro 2U server with Intel Xeon > E5620, 8 gigs of RAM, 1TB HDD and a network interface that uses the > e1000e driver (which is why I couldn't use the SystemImager 4.0.2 > distribution, as that version of BOEL doesn't have the e1000e module > built into the kernel). The golden client kernel version is > > 2.6.38.6-27.fc15.x86_64 > > I've gotten netboot setup to use the standard (BOEL) install > environment. Installation proceeds normally, partitions the HDD, grabs > files through rsync then halts here: > > mount /dev /a/dev -o bind || shellout > Using pci-automod to generate hardware.lst (this may take a long time): > done > sort: standard output: Invalid argument > Detecting systemconfigurator: > --- > FATAL: kernel too old > WARNING: systemconfigurator was not found in the image fedora15 > --- > FATAL: kernel too old > Killing off running processes. > > write_variables > (then the standard "installation failed" message.) > > I cannot use UYOK for some reason as it fails to detect /proc (emits a > "cannot copy /proc/mounts -- is /proc mounted?" message then dies). > > Is there anything I overlooked? > > Thanks! > > Tom Yue ------------------------------------------------------------------------------ What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay _______________________________________________ sisuite-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/sisuite-users
