[Dovecot] 1.0.rc26 released
Dan Price
dp at eng.sun.com
Fri Mar 9 21:23:55 EET 2007
On Thu 08 Mar 2007 at 01:50PM, Dan Price wrote:
> On Thu 08 Mar 2007 at 10:09PM, Timo Sirainen wrote:
> >
> > Do you have login_process_per_connection=yes? If so, they're probably
> > proxying SSL connections.
>
> Ahh. Ok. I didn't have SSL in the pilot. Should I set
> login_process_per_connection=no? I don't know the best practice here,
> I took what I thought was the default setting.
>
> > Hmm. I don't think I've seen this assert-crash before. If it's happening
> > often, compiling without -O2 might help more in debugging it.
>
> Ok, done. I'll see if it happens again.
>
> > Looks like I didn't finish fixing the endianess conversion problems.
> > http://dovecot.org/list/dovecot-cvs/2007-March/007997.html should fix
> > them so that it only complains "Corrupted index cache
> > dovecot.index.cache: indexid changed"
Today I've got another new crash I haven't seen before-- this is
badly impacting the user in question-- since we're crashing on their
inbox. This is running on our unoptimized -g only instance-- as requested
yesterday. Should we go back to rc24?
-dp
Info: imap-login: Login: user=<foobar>, method=PLAIN, rip=XXX.YYY.61.79, lip=AAA.BBB.17.59
Info: IMAP(foobar): Effective uid=27522, gid=10, home=/home/foobar
Info: IMAP(foobar): mbox: data=/home/foobar:INBOX=/var/mail/foobar:INDEX=/home/foobar/Mail/imapd-indices
Info: IMAP(foobar): mbox: root=/home/foobar, index=/home/foobar/Mail/imapd-indices, inbox=/var/mail/foobar
Error: IMAP(foobar): file mbox-sync-update.c: line 458 (mbox_sync_update_header_from): assertion failed: (ctx->mail.uid == 0 || ctx->mail.uid == mail->uid)
Error: IMAP(foobar): Raw backtrace: 0x80bdb2f -> 0x808e51e -> 0x808d56b -> 0x80881f0 -> 0x808a1f2 -> 0x808a660 -> 0x80b21b0 -> 0x8075a8a -> 0x806ba5c -> 0x806bc10 -> 0x806da64 -> 0x806dc59 -> 0x806dd77 -> 0x80c31d9 -> 0x80c2c11 -> 0x8077d2c -> 0x8067b90
Error: child 274367 (imap) killed with signal 6
(gdb) bt full
#0 0xfee505a7 in _lwp_kill () from /lib/libc.so.1
No symbol table info available.
#1 0xfee4d71b in thr_kill () from /lib/libc.so.1
No symbol table info available.
#2 0xfee06502 in raise () from /lib/libc.so.1
No symbol table info available.
#3 0xfede4b8d in abort () from /lib/libc.so.1
No symbol table info available.
#4 0x080bdfc8 in i_internal_panic_handler (fmt=0x0, args=0x8046fe4 "i!\r\bÊ\001") at failures.c:406
backtrace = 0x80eedd8 "0x80bdb2f -> 0x808e51e -> 0x808d56b -> 0x80881f0 -> 0x808a1f2 -> 0x808a660 -> 0x80b21b0 -> 0x8075a8a -> 0x806ba5c -> 0x806bc10 -> 0x806da64 -> 0x806dc59 -> 0x806dd77 -> 0x80c31d9 -> 0x80c2c11 -> 0x807"...
#5 0x080bdb2f in i_panic (format=0x80d7838 "file %s: line %d (%s): assertion failed: (%s)") at failures.c:195
No locals.
#6 0x0808e51e in mbox_sync_update_header_from (ctx=0x80474b0, mail=0x81059c8) at mbox-sync-update.c:458
__PRETTY_FUNCTION__ = "mbox_sync_update_header_from"
#7 0x0808d56b in mbox_sync_rewrite (sync_ctx=0x80477f0, mail_ctx=0x0, end_offset=116254496, move_diff=20, extra_space=The value of variable 'extra_space' is distributed across several
locations, and GDB cannot access its value.
) at mbox-sync-rewrite.c:362
new_mail_ctx = {sync_ctx = 0x80477f0, mail = {uid = 1, idx_seq = 0, keywords = {buffer = 0x0, element_size = 0}, flags = 8 '\b', uid_broken = 0,
from_offset = 0, body_size = 1613, offset = 116250425, space = 0}, seq = 9347, hdr_offset = 116250425, body_offset = 116251545,
header_first_change = 4294967295, header_last_change = 0, header = 0x80fa698, hdr_md5_sum = "Ê\020©:Ô\021Ãî*ÉLñ\\\001\217ä",
content_length = 18446744073709551615, hdr_pos = {1083, 4294967295, 4294967295, 1096, 1108}, parsed_uid = 1, last_uid_updated_value = 0,
last_uid_value_start_pos = 0, have_eoh = 1, need_rewrite = 0, seen_imapbase = 0, pseudo = 0, updated = 0, recent = 0, dirty = 0, imapbase_rewrite = 0,
imapbase_updated = 0}
offset = The value of variable 'offset' is distributed across several
locations, and GDB cannot access its value.
DBX does slightly better:
(dbx) where
[1] __lwp_kill(0x1, 0x6), at 0xfee505a7
[2] raise(0x6), at 0xfee06502
[3] abort(0x8047008, 0x80eedd8, 0x8046fd8, 0x80bdb2f, 0x80d7838, 0x8046fe4), at 0xfede4b8d
=>[4] i_internal_fatal_handler(status = 135100472, fmt = (nil), args = (nil)), line 413 in "failures.c"
[5] i_panic(format = 0x80d7838 "file %s: line %d (%s): assertion failed: (%s)", ...), line 195 in "failures.c"
[6] mbox_sync_update_header_from(ctx = (nil), mail = 0x81059c8), line 458 in "mbox-sync-update.c"
[7] mbox_sync_rewrite(sync_ctx = 0x80477f0, mail_ctx = (nil), end_offset = 116254496ULL, move_diff = 20LL, extra_space = 0, first_seq = 9347U, last_seq = 0), line 362 in "mbox-sync-rewrite.c"
[8] mbox_sync_do(sync_ctx = 0x80477f0, flags = MBOX_SYNC_UNDIRTY), line 854 in "mbox-sync.c"
[9] mbox_sync(mbox = (nil), flags = 0), line 1765 in "mbox-sync.c"
[10] mbox_storage_sync_init(box = (nil), flags = 0), line 1836 in "mbox-sync.c"
[11] mailbox_sync_init(box = 0x80478e8, flags = MAILBOX_SYNC_FLAG_FULL_READ), line 406 in "mail-storage.c"
[12] imap_sync_nonselected(box = 0x8101c10, flags = MAILBOX_SYNC_FLAG_FULL_READ), line 196 in "imap-sync.c"
[13] _cmd_select_full(cmd = 0x80fbd4c, readonly = 0), line 39 in "cmd-select.c"
[14] cmd_select(cmd = 0x80fbd4c), line 92 in "cmd-select.c"
[15] client_handle_input(cmd = 0x80fbd4c), line 332 in "client.c"
[16] client_handle_input(cmd = 0x80fbd4c), line 389 in "client.c"
[17] _client_input(context = 0x80fbd08), line 432 in "client.c"
[18] io_loop_handler_run(ioloop = 0x80f9500), line 199 in "ioloop-poll.c"
[19] io_loop_run(ioloop = (nil)), line 311 in "ioloop.c"
[20] main(argc = 1, argv = 0x8047b18, envp = 0x8047b20), line 287 in "main.c"
(dbx) dump
fmt = (nil)
args = (nil)
status = 135100472
(dbx) up
Current function is i_panic
195 panic_handler(format, args);
(dbx) dump
format = 0x80d7838 "file %s: line %d (%s): assertion failed: (%s)"
args = 0x8047018 "hu^D^Hk\xd5^H^H\xb0t^D^H\xc8Y^P^Hhu^D^H\xd7\xd6^H^HHq^D^H`^N\xfb\xfe\xf8p^D^H\xc0\xa6^O^H\x84w^D^H\x98\xaf\xff\xfe\w^D^H^P^D"
(dbx) up
Current function is mbox_sync_update_header_from
458 i_assert(ctx->mail.uid == 0 || ctx->mail.uid == mail->uid);
(dbx) dump
ctx = (nil)
mail = 0x81059c8
__PRETTY_FUNCTION__ = "mbox_sync_update_header_from"
(dbx) up
Current function is mbox_sync_rewrite
362 mbox_sync_update_header_from(mail_ctx, &mails[idx]);
(dbx) dump
first_nonexpunged = 168U
move_diff = 20LL
ret = 0
mails = 0x81059c8
idx = 0
padding_per_mail = 4U
first_nonexpunged_idx = 0
expunged_space = 0
orig_next_uid = 56593U
sync_ctx = 0x80477f0
extra_space = 0
mail_ctx = (nil)
dest_offset = 0
count = 20U
last_seq = 0
orig_prev_msg_uid = 56592U
next_move_diff = 16ULL
end_offset = 116254496ULL
__PRETTY_FUNCTION__ = "mbox_sync_rewrite"
start_offset = 116250373ULL
offset = 0
first_seq = 9347U
next_end_offset = 116250425ULL
(dbx) up
Current function is mbox_sync_do
854 if (mbox_sync_rewrite(sync_ctx,
(dbx) dump
ret = 134510760
flags = MBOX_SYNC_UNDIRTY
sync_ctx = 0x80477f0
mail_ctx = RECORD
st = (nil)
__PRETTY_FUNCTION__ = "mbox_sync_do"
partial = 0
i = 0
(dbx) up
Current function is mbox_sync
1765 ret = mbox_sync_do(&sync_ctx, flags);
(dbx) dump
leave_dirty = 200U
sync_view = 0x81030d8
ret = 0
changed = 1
mbox = (nil)
flags = 0
sync_ctx = RECORD
delay_writes = 200U
seq = 4294967295U
lock_id = 3U
offset = 18446744073709551615ULL
index_sync_ctx = 0x81030a0
(dbx) up
Current function is mbox_storage_sync_init
1836 ret = mbox_sync(mbox, mbox_sync_flags);
(dbx) dump
ret = 0
mbox = 0x80478e8
flags = 0
mbox_sync_flags = 0
box = (nil)
(dbx) up
Current function is mailbox_sync_init
406 return box->v.sync_init(box, flags);
(dbx) dump
flags = MAILBOX_SYNC_FLAG_FULL_READ
box = 0x80478e8
(dbx) up
Current function is imap_sync_nonselected
196 ctx = mailbox_sync_init(box, flags);
(dbx) dump
ctx = (nil)
flags = MAILBOX_SYNC_FLAG_FULL_READ
box = 0x8101c10
sync_rec = RECORD
status = RECORD
(dbx) up
Current function is _cmd_select_full
39 if (imap_sync_nonselected(box, MAILBOX_SYNC_FLAG_FULL_READ) < 0) {
(dbx) dump
client = 0x80fbd08
storage = 0x80fba20
box = 0x8101c10
cmd = 0x80fbd4c
status = RECORD
readonly = 0
mailbox = 0x80fc020 "Mail/work"
(dbx) up
Current function is cmd_select
92 return _cmd_select_full(cmd, FALSE);
--
Daniel Price - Solaris Kernel Engineering - dp at eng.sun.com - blogs.sun.com/dp
More information about the dovecot
mailing list