kernel panic on 6.3, bin/sh cannot find libc.so.6

[log69]

I posted my issue on the ML too, but I thought it's worth to write it here too.

I have updated my system to 6.3 w/ kernel 2.6.32-279.2.1 (x86_64).

When restarting my machine, it gets kernel panic right at the very top of the screen starting from the first line, saying that /bin/sh cannot find libc.so.6.

The update ran just fine, no error or warning messages shown after all.

I tried reinstalling the kernel and glibc but no help.

Anybody has any idea how to resolve it? I can boot up my system with kernel 279.1.1 just fine (it's a laptop, the rest of my systems I'll keep on 6.2 for a while).

Thanks!

[zxq9]

Do you have any other repos installed? Something may have overwritten glibc.

[log69]

QUOTE (zxq9 @ Aug 10 2012, 06:58 AM)

Do you have any other repos installed? Something may have overwritten glibc.

I have, but it doesn't seem like that:

CODE

$ yum list installed | grep -i glibc glibc.i686                          2.12-1.80.el6_3.3           @sl-security     glibc.x86_64                        2.12-1.80.el6_3.3           @sl-security     glibc-common.x86_64                 2.12-1.80.el6_3.3           @sl-security/6.2 glibc-devel.x86_64                  2.12-1.80.el6_3.3           @sl-security/6.2 glibc-headers.x86_64                2.12-1.80.el6_3.3           @sl-security/6.2

[log69]

Now I'm cooking myself a live spin from 6xrolling with a default Gnome desktop and will try to boot it on my notebook. If everything goes well, then I'll reinstall my system. I played a lot with it since I started to get more familiar with SL half a year ago since 6.2 RC, now I know much better how the cookie crumbles

Anyhow, I'll report back with the results.

[log69]

I've finished creating my live spin based on 6xrolling now. I booted it up in VBox, and then I see that the kernel is still 279.1.1. I've updated another of my machine to 6.3 with the command "yum --releasever=6.3 update", and after the update - which went just fine - that one has kernel 279.1.1 too instead of 279.2.1, which is on my notebook. It's weird.

How can I have on my main system kernel 279.2.1 with a normal update? (I've head only 279.1.1 before the update).

Running a simple "yum update" in the live system as root gives no packages marked for update. How can this be? On my notebook I have this:

CODE

$ yum list installed | grep kernel dracut-kernel.noarch                004-284.el6_3               @sl-fastbugs     kernel.x86_64                       2.6.32-220.23.1.el6         @sl-security/6.2 kernel.x86_64                       2.6.32-279.1.1.el6          @sl-security/6.2 kernel.x86_64                       2.6.32-279.2.1.el6          @sl-fastbugs     kernel-devel.x86_64                 2.6.32-220.23.1.el6         @sl-security/6.2 kernel-devel.x86_64                 2.6.32-279.1.1.el6          @sl-security/6.2 kernel-devel.x86_64                 2.6.32-279.2.1.el6          @sl-fastbugs     kernel-firmware.noarch              2.6.32-279.2.1.el6          @sl-fastbugs     kernel-headers.x86_64               2.6.32-279.2.1.el6          @sl-fastbugs     $ cat /etc/redhat-release Scientific Linux release 6.3 (Carbon)

On the live system just this:

CODE

$ yum list installed | grep kernel dracut-kernel.noarch                004-283.el6 kernel.x86_64                       2.6.32-279.1.1.el6 kernel-firmware.noarch              2.6.32-279.1.1.el6 $ cat /etc/redhat-release Scientific Linux release 6.3 (Carbon)

Ahh, now I see. I use fastbugs on my notebook. So now I'm gonna turn that off. I got bitten by it once, now this was the second time.

[log69]

I ran the live system on my notebook for a while with kernel 279.1.1, and it does the same like on my system: wireless drops connection from time to time and kernel panics after a while. I never had any problem with the kernels before though.

Anyhow, I'll stay with 220.23.x now. I hope TUV will release fixes soon for these issues. Saw similar bug reports.

[AndrewSerk]

Hi log69,

Just thought I would post that I spun 6.3 into a live CD today also. I am not experiencing any issues here but I am not using wireless. I have had the new remix running for a few hours (kvm vm accessed across a network) without issue. I am using the same glibc and kernel that you are having issues with but I spun a slightly slimmed down remix with many standard packages removed and with many linuxTECH.net packages added:

CODE

RemixInTime@RemixInTime ~]$ rpm -q glibc glibc-2.12-1.80.el6_3.3.x86_64 [RemixInTime@RemixInTime ~]$ rpm -q kernel kernel-2.6.32-279.1.1.el6.x86_64 RemixInTime@RemixInTime ~]$ rpm -qa|wc    984     984   33083

Andrew

[log69]

I see this change here:
/lib/modules/2.6.32-220.23.1.el6.x86_64/kernel/drivers/net/wireless/iwlwifi/iwl3945.ko
/lib/modules/2.6.32-279.1.1.el6.x86_64/kernel/drivers/net/wireless/iwlegacy/iwl3945.ko

So they renamed the iwlwifi modul to legacy. I wonder if the solution should be sought around here. No one uses here Intel wireless?

[AndrewSerk]

From a laptop:

CODE

$ lspci|grep Network 0b:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN [Kedron] Network Connection (rev 61) $ lsmod|grep iwl iwl4965               103472  0 iwl_legacy             61040  1 iwl4965 mac80211              185308  2 iwl4965,iwl_legacy cfg80211              158671  3 iwl4965,iwl_legacy,mac80211

No issue here yet but not much uptime here.

[log69]

On my machine, the connection starts dropping from a couple of minutes every time. I think I'll dig up and check the changelog of the kernel that I have this issue with.

Thanks for checking.

[log69]

I removed all 279.2.1 kernel packages with devel and firmware files, then removed 279.1.1 and its pkgs too, then reinstalled both and now my system seems to be working fine with 2.1. No wireless issues or panics noticed.

So I guess 279.2.1 was released in fastbugs just to solve issues with 1.1, just somehow my system was screwed up with the /bin/sh somehow, or the initrd in the pkg or whatever.

[tux99]

QUOTE (log69 @ Aug 11 2012, 04:58 PM)

or the initrd in the pkg or whatever.

It might have been quite possibly the initrd that didn't get regenerated correctly.
(the initrd gets regenerated locally on your PC with the kernel modules for your specific hardware)