[Dovecot] Panic in 1.1.6 file ioloop.c: line 206

[Laurent Blume]

Timo Sirainen a écrit :
> I can't really think of why this would have happened. You wouldn't
> happen to have a core dump available from this? That would be very
> helpful in figuring out why this happened.

I have, here are the pstacks:

$ pstack core.dovecot.7911
core 'core.dovecot.7911' of 7911:       /usr/local/dovecot/sbin/dovecot
-c /usr/local/dovecot/etc/dovecot.conf
 ff1c1adc _lwp_kill (6, 0, 0, ff1a4d94, ffffffff, 6) + 8
 ff140218 abort    (384bc, 1, 4a584, aa1a0, ff1ed2d8, 0) + 110
 00024c28 default_fatal_finish (4, 0, 38be0, ffbffb88, 38400, 55e78) + 68
 00024d90 i_panic  (38be0, 4c8f8, ce, 4c904, 20, 1e7a0) + 2c
 0002851c timeout_get_wait_time (7c, ffbffc00, 0, 3e8, 38800, fffffe82)
+ 15c
 000286ac io_loop_handle_timeouts_real (591b0, 5cc70, 17714, 4dc00,
38c00, 4b878) + 120
 00028728 io_loop_handle_timeouts (591b0, ef, 3d0, 10624c00, e8b250,
3a2c9400) + 10
 00028de0 io_loop_handler_run (591b0, 1, 0, 38c00, 0, 5cbd0) + 60
 0002877c io_loop_run (591b0, 80000000, 80, 0, 80000000, 80000000) + 3c
 0001f56c main     (36000, 4d000, 0, 4dc00, 0, 0) + 4f0
 00015c40 _start   (0, 0, 0, 0, 0, 0) + 108

$ pstack core.imap.7944
core 'core.imap.7944' of 7944:  imap
 ff1c1adc _lwp_kill (6, 0, 0, ff1a4d94, ffffffff, 6) + 8
 ff140218 abort    (e0df0, 1, f3f88, aa1a0, ff1ed2d8, 0) + 110
 000b4784 default_fatal_finish (4, 0, e1d58, ffbff818, e0c00, f5cbc) + 68
 000b48ec i_panic  (e1d58, 10c000, ce, 10c00c, a0, b4f00) + 2c
 000bbfe0 timeout_get_wait_time (26c, ffbff890, 0, 1388, e1c00,
ffffff73) + 15c
 000bc170 io_loop_handle_timeouts_real (11b318, 115f80, 6f6d4, 10d000,
e1c00, f5cfc) + 120
 000bc1ec io_loop_handle_timeouts (11b318, 3, 1388, 10624c00, 0, 0) + 10
 000bc8a4 io_loop_handler_run (11b318, 1, 0, e1c00, 80000000, 115aa0) + 60
 000bc240 io_loop_run (11b318, 80000000, 80, 0, 80000000, 80000000) + 3c
 00038a10 main     (1, ffbffa94, ffbffa9c, f4400, 10d000, 10d000) + a0
 000215a8 _start   (0, 0, 0, 0, 0, 0) + 108


Does it help?


> But anyway I did some changes to that code, maybe it'll give a different
> assert-crash the next time:
> http://hg.dovecot.org/dovecot-1.1/rev/75f36619846f

Thanks, I'll try to reproduce the problem and apply those fixes,

Laurent
-- 
/ Leader de Projet & Communauté    | I'm working, but not speaking for
\ G11N   http://fr.opensolaris.org | Bull Services http://www.bull.com
/ FOSUG  http://guses.org          |