[Dovecot] Solaris 10 / 1.1.beta15 imap cores
Hi,
I've been using 1.0 but moved to beta to see if it works before release. Both b14 and b15 dies sometimes when accessing folders. Last time I fixed this by moving one message from folder to another and then back. Next I'm trying to build this with debugging enabled.
Tomppa
Feb 14 11:52:33 dovecot: [ID 107833 mail.crit] IMAP(tomppa): file message-parser.c: line 674: assertion failed: (ctx->part->physical_pos >= ctx->input->v_offset) Feb 14 11:52:34 genunix: [ID 603404 kern.notice] NOTICE: core_log: imap[5995] core dumped: /v/tmp/core.imap.5995 Feb 14 11:52:40 dovecot: [ID 107833 mail.crit] IMAP(tomppa): file message-parser.c: line 674: assertion failed: (ctx->part->physical_pos >= ctx->input->v_offset) Feb 14 11:52:40 genunix: [ID 603404 kern.notice] NOTICE: core_log: imap[5997] core dumped: /v/tmp/core.imap.5997 Feb 14 11:52:47 dovecot: [ID 107833 mail.crit] IMAP(tomppa): file message-parser.c: line 674: assertion failed: (ctx->part->physical_pos >= ctx->input->v_offset) Feb 14 11:52:47 genunix: [ID 603404 kern.notice] NOTICE: core_log: imap[5999] core dumped: /v/tmp/core.imap.5999
)# dbx /usr/local/libexec/dovecot/imap /v/tmp/core.imap.5999 For information about new features see
help changes' To remove this message, put
dbxenv suppress_startup_message 7.6' in your .dbxrc Reading imap core file header read successfully Reading ld.so.1 Reading librt.so.1 Reading libnsl.so.1 Reading libsocket.so.1 Reading libsendfile.so.1 Reading libc.so.1 Reading libaio.so.1 Reading libmd.so.1 Reading libc_psr.so.1 Reading 8859-1%UTF-8.so program terminated by signal ABRT (Abort) 0xff1c5c14: __lwp_kill+0x0008: bcc,a,pt %icc,__lwp_kill+0x18 ! 0xff1c5c24 (dbx) where
=>[1] __lwp_kill(0x0, 0x6, 0x0, 0x6, 0xfc00, 0x0), at 0xff1c5c14 [2] raise(0x6, 0x0, 0x0, 0xff1a8c20, 0xffffffff, 0x6), at 0xff164c44 [3] abort(0x2, 0x1, 0xe83c4, 0xad324, 0xff1f1318, 0x0), at 0xff1410f8 [4] 0xaa4f0(0x2, 0x100400, 0xd5400, 0x100ba0, 0xea000, 0xea078), at 0xaa4f0 [5] i_panic(0xd4a6c, 0xea000, 0x2a2, 0xfeb88, 0xea16c, 0xaa464), at 0xa9de4 [6] 0xa2ec4(0x157170, 0xffbff544, 0xd4800, 0x0, 0x1380e0, 0xfe800), at 0xa2ec4 [7] message_parser_parse_next_block(0x157170, 0xffbff544, 0x1, 0x0, 0xfe800, 0xa2d4c), at 0xa3064 [8] 0xa4554(0x151158, 0x14fc88, 0x138090, 0xffbff534, 0x80000000, 0xd4cc4), at 0xa4554 [9] message_search_msg(0x151158, 0x14fc88, 0x137ff0, 0x14fc60, 0xe0, 0x60), at 0xa45d4 [10] 0x67a40(0x10c588, 0x10, 0x10c588, 0x14fb38, 0x151158, 0xffbff6e4), at 0x67a40 [11] mail_search_args_foreach(0x10c588, 0x679a4, 0xffbff6e4, 0x1, 0x1, 0xffffffff), at 0x6f688 [12] 0x67d68(0x10c550, 0x137b68, 0x67800, 0xffffffff, 0xffffffff, 0x14fc88), at 0x67d68 [13] index_storage_search_next_nonblock(0x137b68, 0x137be8, 0x1, 0xffbff81f, 0xea000, 0x113b28), at 0x69368 [14] 0x26008(0x10c500, 0x40000000, 0x2, 0x10c5c8, 0x80000000, 0xc2000), at 0x26008 [15] 0x262c8(0x10c500, 0x100550, 0x1, 0xea000, 0xea058, 0x10c2a0), at 0x262c8 [16] 0xb1c10(0x109518, 0x10a610, 0x3, 0x3e8, 0xea0b8, 0xd64f8), at 0xb1c10 [17] io_loop_handle_timeouts(0x109518, 0x3, 0x0, 0x10624c00, 0x0, 0x0), at 0xb1c70 [18] io_loop_handler_run(0x109518, 0x0, 0x2, 0x29414, 0xd6400, 0x10a5b0), at 0xb2380 [19] io_loop_run(0x80, 0x0, 0x80000000, 0x80000000, 0x10a5b0, 0x109518), at 0xb1cc4 [20] main(0x100800, 0x10c2a0, 0xe8c00, 0x10000, 0x100978, 0x1008c4), at 0x330ac (dbx)
# /usr/local/sbin/dovecot -n # 1.1.beta15: /usr/local/etc/dovecot.conf Warning: fd limit 256 is lower than what Dovecot can use under full load (more than 640). Either grow the limit or change login_max_processes_count and max_mail_processes settings base_dir: /var/run/dovecot/ ssl_cert_file: /etc/mail/certs/imapd.pem ssl_key_file: /etc/mail/certs/key.pem disable_plaintext_auth: no login_dir: /var/run/dovecot//login login_executable: /usr/local/libexec/dovecot/imap-login first_valid_uid: 100 mail_location: mbox:~/Posti/:INBOX=/var/mail/%u:INDEX=/var/dovecot/indexes/%u auth default: passdb: driver: pam userdb: driver: passwd
On Thu, 2008-02-14 at 13:07 +0200, Tomi Vainio wrote:
Feb 14 11:52:33 dovecot: [ID 107833 mail.crit] IMAP(tomppa): file message-parser.c: line 674: assertion failed: (ctx->part->physical_pos >= ctx->input->v_offset)
Any idea how I could reproduce this? I've been running my IMAP stress test for a while now waiting for this assert to fail, but so far it hasn't. Maybe it needs some special kind of a mail or some other specific condition.
Anyway I did some fixes, do these change the crash to an error message?
http://hg.dovecot.org/dovecot-1.1/rev/a6c066f50877 http://hg.dovecot.org/dovecot-1.1/rev/8cfa61f98e32 http://hg.dovecot.org/dovecot-1.1/rev/289765861d66
Timo Sirainen wrote:
On Thu, 2008-02-14 at 13:07 +0200, Tomi Vainio wrote:
Feb 14 11:52:33 dovecot: [ID 107833 mail.crit] IMAP(tomppa): file message-parser.c: line 674: assertion failed: (ctx->part->physical_pos >= ctx->input->v_offset)
Any idea how I could reproduce this? I've been running my IMAP stress test for a while now waiting for this assert to fail, but so far it hasn't. Maybe it needs some special kind of a mail or some other specific condition.
Anyway I did some fixes, do these change the crash to an error message?
http://hg.dovecot.org/dovecot-1.1/rev/a6c066f50877 http://hg.dovecot.org/dovecot-1.1/rev/8cfa61f98e32 http://hg.dovecot.org/dovecot-1.1/rev/289765861d66
This problem don't occur very often. So far only seven times since Jan 31 and everytime I have done something which fixes it magically. Yesterday it was a mail which had only base64 encoded data in it. I was able to move that mail to another folder and back using thunderbird. I've no idea why have the problem has appeared more frequently during the last couple of days. Here are the timestamps when it happened and first and last log entry. Altogether 315 cores so far. I started to use 1.1 b14 Jan 23 so it took a while before it happened. I'm the only one using this environment.
Jan 31 14:16:10 [kern.notice] NOTICE: core_log: imap[3416] core dumped: /v/tmp/core.imap.3416 Jan 31 14:16:58 [kern.notice] NOTICE: core_log: imap[13449] core dumped: /v/tmp/core.imap.13449
Feb 4 09:24:03 [kern.notice] NOTICE: core_log: imap[20972] core dumped: /v/tmp/core.imap.20972 Feb 4 09:25:01 [kern.notice] NOTICE: core_log: imap[14911] core dumped: /v/tmp/core.imap.14911
Feb 5 18:37:32 [kern.notice] NOTICE: core_log: imap[16454] core dumped: /v/tmp/core.imap.16454 Feb 5 18:37:39 [kern.notice] NOTICE: core_log: imap[16679] core dumped: /v/tmp/core.imap.16679
Feb 12 21:36:06 [kern.notice] NOTICE: core_log: imap[7172] core dumped: /v/tmp/core.imap.7172 Feb 12 21:40:52 [kern.notice] NOTICE: core_log: imap[7731] core dumped: /v/tmp/core.imap.7731
Feb 13 11:52:03 [kern.notice] NOTICE: core_log: imap[4618] core dumped: /v/tmp/core.imap.4618 Feb 13 11:52:43 [kern.notice] NOTICE: core_log: imap[620] core dumped: /v/tmp/core.imap.620
Feb 13 12:50:23 [kern.notice] NOTICE: core_log: imap[648] core dumped: /v/tmp/core.imap.648 Feb 13 12:50:41 [kern.notice] NOTICE: core_log: imap[759] core dumped: /v/tmp/core.imap.759
Feb 14 11:43:07 [kern.notice] NOTICE: core_log: imap[5528] core dumped: /v/tmp/core.imap.5528 Feb 14 11:52:47 [kern.notice] NOTICE: core_log: imap[5999] core dumped: /v/tmp/core.imap.5999
Tomppa
Timo Sirainen wrote:
On Thu, 2008-02-14 at 13:07 +0200, Tomi Vainio wrote:
Feb 14 11:52:33 dovecot: [ID 107833 mail.crit] IMAP(tomppa): file message-parser.c: line 674: assertion failed: (ctx->part->physical_pos >= ctx->input->v_offset)
Any idea how I could reproduce this? I've been running my IMAP stress test for a while now waiting for this assert to fail, but so far it hasn't. Maybe it needs some special kind of a mail or some other specific condition.
Anyway I did some fixes, do these change the crash to an error message?
http://hg.dovecot.org/dovecot-1.1/rev/a6c066f50877 http://hg.dovecot.org/dovecot-1.1/rev/8cfa61f98e32 http://hg.dovecot.org/dovecot-1.1/rev/289765861d66
I've build debug version of imap with these changes. I've been running it two hours without seeing cores when doing searches on the same folders which caused problems yesterday.
Tomppa
participants (3)
-
Timo Sirainen
-
Tomi Vainio
-
Tomi Vainio