Hide Forgot
Description of problem: imap/dovecot crash after segfault in dovecot 2.0.9 Version-Release number of selected component (if applicable): Version : 2.0.9 Release : 2.el6 How reproducible: Currently not possible. Actual results: /var/log/maillog Jun 9 00:12:34 mail dovecot: master: Error: service(imap): child 8970 killed with signal 11 (core dumps disabled) Jun 9 00:12:34 mail dovecot: imap-login: Login: user=<username>, method=PLAIN, rip=IP, lip=IP, mpid=8987, TLS Jun 9 00:12:34 mail dovecot: master: Error: service(imap): child 8987 killed with signal 11 (core dumps disabled) /var/log/messages Jun 9 00:12:34 mail kernel: imap[8970]: segfault at 0 ip 00007fe4b2fe432d sp 00007fff29383740 error 4 in libdovecot-storage.so.0.0.0[7fe4b2fb8000+d1000] Jun 9 00:12:34 mail kernel: imap[8987]: segfault at 0 ip 00007f949de1532d sp 00007fff120721b0 error 4 in libdovecot-storage.so.0.0.0[7f949dde9000+d1000]
Are you able to get backtrace for this crash? If you have abrt installed it should catch it, so it should be easy (using abrt-gui you can get backtrace or you can create backtrace manually from core file located in /var/spool/abrt/...) Also please paste output of doveconf -n
Hi Michal, abrt is not installed. Sorry, no backtrace available. The bug does not seem to occur often. (therefore not reproducible) Perhaps the bug has been fixed with a newer version. See http://dovecot.org/pipermail/dovecot/2011-June/059643.html
Hi, I'll still need doveconf -n output (so *maybe* I can find some related bug fixed), but without backtrace or reproducer I can't do much more, as Timo already told you on mailing list. Also we don't rebase packages in RHEL so often without justification, so "perhaps fixed" won't help here. I can built test package for you if you want to test if this bug is fixed in newer version. But, because dovecot was not approved for 6.2, it'd help filling this issue though regular support page ( https://access.redhat.com/support/ ) to get higher priority (bugzilla is just bug tracking tool, bugs should be reported through support channels).
Hi Michal, I sent you an email with further information.
Hi, I did not find any similar bug and without backtrace or reproducer I can just built test package if you want to test if latest dovecot version works
Hi Michal, Thank you. Test packages would be a good idea.
I've uploaded *test* packages here: http://mhlavink.fedorapeople.org/bz712584/
Thank you for your efforts. I'll try the packages and then inform you if everything is okay.
This request was evaluated by Red Hat Product Management for inclusion in the current release of Red Hat Enterprise Linux. Because the affected component is not scheduled to be updated in the current release, Red Hat is unfortunately unable to address this request at this time. Red Hat invites you to ask your support representative to propose this request, if appropriate and relevant, in the next release of Red Hat Enterprise Linux. If you would like it considered as an exception in the current release, please ask your support representative.
Hi, Small update on this: In 6 months the error occurred only once. The bug was not reproducible. There was a BIOS update for IBM servers which resolves the following microcode error: - Fixed : Intel microcode fix for a potential memory addressing error in 64-bit operating mode; it could cause page fault, data loss, data corruption or other unpredictable system behavior. So I am assuming that it was not a Dovecot bug. Best regards, Morten
> So I am assuming that it was not a Dovecot bug. OK, I'm closing this bug. If you see it again, feel free to reopen