Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: exim: dev

[Bug 1178] queue runner is using huge amounts of memory

 

 

exim dev RSS feed   Index | Next | Previous | View Threaded


graeme at graemef

Jan 9, 2012, 4:28 AM

Post #1 of 5 (679 views)
Permalink
[Bug 1178] queue runner is using huge amounts of memory

------- You are receiving this mail because: -------
You are on the CC list for the bug.

http://bugs.exim.org/show_bug.cgi?id=1178




--- Comment #11 from Graeme Fowler <graeme [at] graemef> 2012-01-09 12:28:55 ---
Thanks, Yuri.

From an initial parse of that strace output, we see some 35 or so seconds of
the end-to-end life of a queue runner - during which it is spawning processes
at quite a rate.

The parent of the majority of the spawned processes is PID 8174. In the
timeframe of this trace, it spawns 2412 child processes (cloned).

Each one is iterating over the same set of addresses.

All that said, there are no OOM conditions in this strace - no ENOMEM errors
are recorded anywhere, and all the cloned children seem to exit gracefully. At
the end of the trace, PIDs 8174 and 8149 (the invocation PID) exit with a
SIGTERM, indicating most likely that you hit CTRL-C on the keyboard. It would
have been interesting to see it go to completion.

The problematic message ID (Exim's internal queue ID) is 1Riooa-0006r5-2f.

I'm hedging towards a backing DBM error here, but that's a long shot. If I get
a bit more time to look later today, I will.


--
Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email

--
## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##


admins at fused

Jan 9, 2012, 1:52 AM

Post #2 of 5 (684 views)
Permalink
[Bug 1178] queue runner is using huge amounts of memory [In reply to]

------- You are receiving this mail because: -------
You are on the CC list for the bug.

http://bugs.exim.org/show_bug.cgi?id=1178




--- Comment #8 from Yuri Arabadji <admins [at] fused> 2012-01-09 09:52:55 ---
Alright, it started happening again. So when I view the message log file, I see
the followin:

================================
2012-01-05 09:32:46 Received from maraudingkb8520 [at] usps H=([58.211.173.250])
[58.211.173.250] P=smtp S=294981
id=CRPC3M54MCQ1Z4359OTRC8A1X9VB82R9S8HARDAMLQO6IN4083793767923 [at] ku8-dc
2012-01-06 15:00:03 mary [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is mary [at] xx): failure to transfer data from subprocess: status=5e959ba0
readerror='Cannot allocate memory'
2012-01-06 15:00:03 dave [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is dave [at] xx): failure to transfer data from subprocess: status=5e95a520
readerror='Cannot allocate memory'
2012-01-06 15:00:03 diana [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is diana [at] xx): failure to transfer data from subprocess: status=5e95d3b8
readerror='Cannot allocate memory'
2012-01-06 15:00:03 shaun [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is shaun [at] xx): failure to transfer data from subprocess: status=5e95dd80
readerror='Cannot allocate memory'
2012-01-06 15:00:03 nico [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is nico [at] xx): failure to transfer data from subprocess: status=5e95e710
readerror='Cannot allocate memory'
2012-01-06 16:07:12 dave [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is dave [at] xx): failure to transfer data from subprocess: status=601cc860
readerror='Cannot allocate memory'
2012-01-06 16:10:41 diana [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is diana [at] xx): failure to transfer data from subprocess: status=60313980
readerror='Cannot allocate memory'
2012-01-06 16:10:42 nico [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is nico [at] xx): failure to transfer data from subprocess: status=60315f60
readerror='Cannot allocate memory'
2012-01-06 16:10:42 mary [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is mary [at] xx): failure to transfer data from subprocess: status=603168e0
readerror='Cannot allocate memory'
2012-01-06 16:10:42 dave [at] xx <pub [at] xx>
R=has_alias_but_no_mailbox_discarded_to_prevent_loop defer (-1): internal
problem in has_alias_but_no_mailbox_discarded_to_prevent_loop router (recipient
is dave [at] xx): failure to transfer data from subprocess: status=60317260
readerror='Cannot allocate memory'
================================

Strace of this very process is attached. Current situation with RAM:
# free -m
total used free shared buffers cached
Mem: 12013 10133 1879 0 452 5540
-/+ buffers/cache: 4140 7872
Swap: 4094 0 4094


--
Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email

--
## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##


admins at fused

Jan 9, 2012, 1:53 AM

Post #3 of 5 (648 views)
Permalink
[Bug 1178] queue runner is using huge amounts of memory [In reply to]

------- You are receiving this mail because: -------
You are on the CC list for the bug.

http://bugs.exim.org/show_bug.cgi?id=1178




--- Comment #9 from Yuri Arabadji <admins [at] fused> 2012-01-09 09:53:36 ---
Created an attachment (id=530)
--> (http://bugs.exim.org/attachment.cgi?id=530)
exim in a loop


--
Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email

--
## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##


admins at fused

Jan 9, 2012, 2:04 AM

Post #4 of 5 (648 views)
Permalink
[Bug 1178] queue runner is using huge amounts of memory [In reply to]

------- You are receiving this mail because: -------
You are on the CC list for the bug.

http://bugs.exim.org/show_bug.cgi?id=1178




--- Comment #10 from Yuri Arabadji <admins [at] fused> 2012-01-09 10:04:20 ---
The file you are trying to attach is 4668 kilobytes (KB) in size. Non-patch
attachments cannot be more than 1000 KB.

Thus download the strace -fFv -tt /usr/sbin/exim -q here:
http://s7.fused.com/exim.queue.run.bz2

Thanks!


--
Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email

--
## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##


admins at fused

Jan 23, 2012, 3:48 AM

Post #5 of 5 (603 views)
Permalink
[Bug 1178] queue runner is using huge amounts of memory [In reply to]

------- You are receiving this mail because: -------
You are on the CC list for the bug.

http://bugs.exim.org/show_bug.cgi?id=1178




--- Comment #12 from Yuri Arabadji <admins [at] fused> 2012-01-23 11:48:02 ---
Created an attachment (id=538)
--> (http://bugs.exim.org/attachment.cgi?id=538)
strace of -Mc run

Current situation:
# ps axwwl|grep [e]xim
4 47 11021 1 27 12 64352 2012 - SNs ? 1:16
/usr/sbin/exim -bd -q60m
4 0 20466 1 27 12 4348992 3428200 - SN ? 2012:59
/usr/sbin/exim -Mc 1RoMKM-0005Jg-92
4 0 20900 1 28 12 4350132 3737908 - RN ? 2011:53
/usr/sbin/exim -Mc 1RoMKM-0005Ji-Pj

AFAIK that was my invocation of -Mc to deliver all messages in the queue. So it
stuck on 2 of them, both destined for non-existent recipients, which, erm, were
there, but somehow passwd file and all account info got deleted.

I have attached "strace -fFv -tt -p 20900" output. Even though the config files
may be missing, the process shouldn't grow to 4GB RSS.


--
Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email

--
## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##

exim dev RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.