Description of Problem: Following a large number of IO errors due to a corrupt fs, ext3 paniced with Assertion failure in journal_forget_Rb6ebb66a() at transaction.c:1224: "!jh->b_committed_data" kernel BUG at transaction.c:1224! Version-Release number of selected component (if applicable): 2.4.18-14 athlon
Created attachment 85004 [details] Kernel log of errors
Saw the same problem 3 times on 2 different servers the last 2 days. One of them is connected to a serial console. Attaching the errors logged (note that there's a problem with newline characters, therefore the log looks somewhat strange ...). This is RHAT 7.3, kernel 2.4.18-17.7.xsmp (2 1GHz processor). (SMP prob?) Only in one case previous ext3 errors were logged. I now rebooted and forced a fsck. There were unexpected inconsitencies, I had to manually run fsck. Any workaround for now (or should I be save for some time having now once repaired the inconsistencies)? Changing severity to high.
Created attachment 88963 [details] kernel error log 1
Created attachment 88964 [details] kernel error log 2
Hm, is this a dup of bug #71223 ?
This week another 2 servers crashed because of this. Doesn't this happen to others, too? This is definitely a no-go for servers ... Any chance this was fixed in the latest kernel update?
It wasn't fixed, as I haven't been able to trace down the exact code path causing this. But recent ext3 does have a new mechanism to allow us to downgrade the severity of assertion failures where there's a chance of triggering the failures due to hardware corruptions or data corruptions, rather than pure software faults, so there's an infrastructure there which will let me relax this message from a panic later on.
patch included in U4 erratum candidate. moving to modified.
An errata has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2004-105.html