https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c1 --- Comment #1 from Andreas Haumer 2012-08-09 14:56:44 UTC --- Created an attachment (id=501817) --> (http://bugzilla.novell.com/attachment.cgi?id=501817) The actual kernel panic message -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c3 Andreas Haumer changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|andreas@xss.co.at | --- Comment #3 from Andreas Haumer 2012-08-10 08:49:58 UTC --- I think I found the cause of the problem. In fact, I think there are two bugs (should I open separate bugzilla entries?) The hwclock being wrong triggers a fsck problem (fsck thinks there is a file system inconsistency because the time stamp of the root-fs superblock is in the future, compared to the time reported by the hardware clock). The system then wants to go into manual repair mode. But it can't start the shell because it doesn't find libkmod.so.2 And it doesn't find this library as this library is placed in /usr/lib64 and on this test system /usr is on a separate logical volume which is not mounted at that stage! Thus: init dies, kernel panic! (see screenshot) I would expect: 1) A wrong hardware clock should *not* bring the system into manual repair mode! This can happen quite easily and it would be a big problem if the system won't boot just because of such a simple issue. This is bug #1 2) A separate /usr filesystem should *not* prevent the system from entering manual repair mode (with root-fs mounted only)! A separate /usr filesystem should be a legal and supported configuration. On all installations we use (server and client), we have a separate, read-only mounted /usr filesystem. This has proven to be very useful, OpenSuse 12.2 should not change this! All files (programs, shared libraries, configs) needed in this mode should be placed on the root-fs! This is bug #2 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c5 Jeff Mahoney changed: What |Removed |Added ---------------------------------------------------------------------------- Component|Kernel |Basesystem AssignedTo|kernel-maintainers@forge.pr |mmarek@suse.com |ovo.novell.com | --- Comment #5 from Jeff Mahoney 2012-08-14 09:46:19 EDT --- My initrd has that library but I'm using mkinitrd-2.7.1-1.1.x86_64 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c7 --- Comment #7 from Andreas Haumer 2012-08-16 11:04:21 UTC --- Created an attachment (id=502509) --> (http://bugzilla.novell.com/attachment.cgi?id=502509) Reproducing the problem on another test machine I just tested this on another machine (Desktop PC, OpenSuse 12.2 RC2 + current patches) and can reproduce the boot problem. The machine does not panic, though, but "just" wants to enter manual repair mode if the hardware clock is set wrong. Doing so, it encounters all kinds of "command not found" and "cannot open shared object file" errors. This is with mkinitrd-2.7.0-62.5.1.x86_64 To me, fsck complaining and stopping the boot process, entering manual repair mode, just because the hardware clock is wrong, is the most serious problem here. In fact I would consider this a showstopper (though it seems other linux distributions do have this problem, too). But this is an admin's nightmare: think of a server located somewhere in a server room, rebooted after one year and it turns out it has a bad CMOS battery. The server could be thousands of kilometers away and fsck wants to force me to get in front of the console just because the hardware clock is set backwards? Or think of a Linux workstation or notebook of some simple user. How can I as admin explain to my users that they have to bring me their equipment (or I have to come to them) just because the computer's clock is wrong? This is ridiculous! What is the rationale behind this behaviour of fsck, anyway? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c9 --- Comment #9 from Andreas Haumer 2012-08-16 12:00:08 UTC --- Found the following related info: Ted Ts'o at Google+: https://plus.google.com/117091380454742934025/posts/3FRcnDqa4iy Fedora's Bugzilla entry for exactly the same problem: https://bugzilla.redhat.com/show_bug.cgi?id=748920 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c11 --- Comment #11 from Ismail Donmez 2012-08-24 09:39:43 UTC --- SR# 131510 handles the e2fsprogs part, we'll also need a mkinitrd modification to include /etc/e2fsck.conf in the initrd image. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c13 Mark Post changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED CC| |mpost@suse.com InfoProvider|mmarek@suse.com | --- Comment #13 from Mark Post 2013-08-06 21:23:04 UTC --- Because /etc/e2fsck.conf is not being included in the initrd, this problem is still happening on openSUSE 12.3. (At least I'm assuming that is the cause, since the system has /etc/e2fsck.conf with the broken_system_clock = 1 in it, but that file isn't in the initrd.) It's been a year, and SLE 12 is coming, so we should probably fix this. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c15 Olaf Hering changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #15 from Olaf Hering 2013-08-08 12:32:09 CEST --- The file is copied correctly with the change above. Closing. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=775268 https://bugzilla.novell.com/show_bug.cgi?id=775268#c17 --- Comment #17 from Swamp Workflow Management 2013-09-09 12:06:08 UTC --- openSUSE-RU-2013:1416-1: An update that has 10 recommended fixes can now be installed. Category: recommended (moderate) Bug References: 755642,775268,784613,787075,801984,804398,805059,809742,809812,814540 CVE References: Sources used: openSUSE 12.3 (src): mkinitrd-2.7.2-3.5.1 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.