LTC Owner is: thinh.com LTC Originator is: thinh.com Problem description: CellBlade CB1 keeps rebooting after install FC6-Test3 (installation details are in LTC bug 27222/ RedHat Bugzilla 206509.) Hardware Environment Machine type (p650, x235, SF2, etc.): BladeCenter QS20 Is this reproducible? YES If so, how long does it (did it) take to reproduce it? 1 minutes Describe the steps: - install the machine, and watch it reboots Additional information: The yaboot.conf sh-3.1# cat yaboot.conf # yaboot.conf generated by anaconda boot=/dev/hda1 init-message=Welcome to Fedora Core!\nHit <TAB> for boot options partition=2 timeout=80 install=/usr/lib/yaboot/yaboot delay=5 enablecdboot enableofboot enablenetboot nonvram fstype=raw image=/vmlinuz-2.6.17-1.2630.fc6 label=linux read-only root=/dev/hda5 append="console=hvc0 rhgb quiet" there is no initrd #ls boot System.map-2.6.17-1.2630.fc6 lost+found config-2.6.17-1.2630.fc6 symsets-2.6.17-1.2630.fc6.tar.gz etc symvers-2.6.17-1.2630.fc6.gz grub vmlinuz-2.6.17-1.2630.fc6 I got it up. Starting smartd: [ OK ] Fedora Core release 5.92 (FC6 Test3) Kernel 2.6.17-1.2630.fc6 on an ppc64 login: root Password: # uname -a Linux localhost 2.6.17-1.2630.fc6 #1 SMP Wed Sep 6 17:17:26 EDT 2006 ppc64 ppcx this is the issue without initrd as Arnd mentioned in the email. The work-around: 1- do netboot as normal vnc installation steps, until anaconda start vncserver and gives back the shell prompt 2- mount the '/' and '/boot' partitions from local hard drive. in my case: mount /dev/hda5 /oldtmp mount /dev/hda2 /oldtmp/boot 3- do chroot to "/oldtmp" chroot /oldtmp 4- run mkinitrd to create initrd image mkinitrd initrd-2.6.17-1.2630.fc6 2.6.17-1.2630.fc6 5- edit yaboot.conf to add initrd # yaboot.conf generated by anaconda boot=/dev/hda1 init-message=Welcome to Fedora Core!\nHit <TAB> for boot options partition=2 timeout=80 install=/usr/lib/yaboot/yaboot delay=5 enablecdboot enableofboot enablenetboot nonvram fstype=raw image=/vmlinuz-2.6.17-1.2630.fc6 label=linux read-only initrd=/initrd-2.6.17-1.2630.fc6 root=/dev/hda5 append="console=hvc0 rhgb quiet" 6- restart the blade
Created attachment 136300 [details] boot_messages
----- Additional Comments From thinh.com (prefers email at th2tran.com) 2006-09-14 19:04 EDT ------- for the work-around: - Don't reboot the machine after installation completed, just go to the shell prompt and do create the initrd image from step #2 above. Regards, Thinh Tran
Can you provide /root/install.log?
Created attachment 136357 [details] requested install log Here is an install.log file from a Cell blade exhibiting this behaviour.
Created attachment 136358 [details] install.log
----- Additional Comments From robbiew.com 2006-09-15 09:48 EDT ------- requested install log
----- Additional Comments From arndb.com (prefers email at arnd.bergmann.com) 2006-09-15 09:54 EDT ------- The bug is probably related to this snipped from the install log WARNING: /lib/modules/2.6.17-1.2630.fc6/kernel/fs/cachefiles/cachefiles.ko needs unknown symbol copy_4K_page /sbin/mkinitrd: line 282: /sbin/nash: No such file or directory /sbin/mkinitrd: line 282: /sbin/nash: No such file or directory /sbin/mkinitrd: line 282: /sbin/nash: No such file or directory /sbin/mkinitrd: line 282: /sbin/nash: No such file or directory no temporary directory could be found. mkinitrd failed error: %post(kernel-2.6.17-1.2630.fc6.ppc64) scriptlet failed, exit status 1 Maybe a dependency in the kernel package to nash is missing?
Created attachment 136359 [details] install.log
----- Additional Comments From thinh.com (prefers email at th2tran.com) 2006-09-15 10:02 EDT ------- /root/install.log mkinitrd fails: .... Installing kernel - 2.6.17-1.2630.fc6.ppc64 WARNING: /lib/modules/2.6.17-1.2630.fc6/kernel/fs/cachefiles/cachefiles.ko needs unknown symbol copy_4K_page /sbin/mkinitrd: line 282: /sbin/nash: No such file or directory /sbin/mkinitrd: line 282: /sbin/nash: No such file or directory /sbin/mkinitrd: line 282: /sbin/nash: No such file or directory /sbin/mkinitrd: line 282: /sbin/nash: No such file or directory no temporary directory could be found. mkinitrd failed error: %post(kernel-2.6.17-1.2630.fc6.ppc64) scriptlet failed, exit status 1 Installing firstboot-tui - 1.4.19-1.noarch ....
this bug seems to be the same as https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=205194.
----- Additional Comments From thinh.com (prefers email at th2tran.com) 2006-09-15 10:12 EDT ------- Redhat, Sorry for sending duplicated data.
Hmm... what's happening is that we're having things sorted such that mkinitrd.ppc64 is installed later than the kernel and rpm is only putting down the ppc64 binary
two points: - mkinitrd.ppc is installed just before the kernel, mkinitrd.ppc64 is installed much later. Why do we install them both in the first place? why doesn't mkinitrd.ppc do the right thing? - the last three changelog entries in mkinitrd.spec refer to bug 205194, but none of them actually fixed it...
Bingo! This same problem occurs on a x86_64 system installing the 64 bit fc6test3. Doing a 32 bit install works ok. In both cases there is a 32 bit and a 64 bit mkinitrd being installed but the 64 bit one is installed AFTER the kernel is installed.
If you indicate here when this is fixed in rawhide, I will take a shot at testing the fix.
changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |gerhard.stenzel.com ------- Additional Comments From gerhard.stenzel.com 2006-09-19 11:09 EDT ------- The same problem (and workaround) occur also on a JS20
*** This bug has been marked as a duplicate of 206453 ***