Mailing List Archive: exim: users

exim delivery process hang at 100% cpu

Index | Next | Previous | View Threaded

david at ols

Apr 12, 2012, 1:23 AM

Post #1 of 6 (479 views)
Permalink

exim delivery process hang at 100% cpu

Hi

Today exim eat all server cpu to the point that i was unable to
login as root. Once recovered i see one exim process eating 100%
cpu when delivering a message to a remote host

[log]# top
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
2855 exim 20 0 42168 2412 1232 R 100 0.0 15:44.75 exim

[log]# /usr/local/exim/bin/exiwhat
2853 delivering 1SIDML-0008Ls-OK: waiting for a remote delivery
subprocess to finish
2855 delivering 1SIDML-0008Ls-OK to fltr-in1.mail.dreamhost.com
[208.97.132.73] (kvmlcq3 [at] dub)

this process never ended and i have to kill it but in the next
queue run the same message was processed correctly

--
Best regards ...

----------------------------------------------------------------
David Saez
On-Line Services 2000 S.L.
http://www.ols.es
----------------------------------------------------------------

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

david at ols

May 31, 2012, 7:52 AM

Post #2 of 6 (413 views)
Permalink

Re: exim delivery process hang at 100% cpu [In reply to]

Hi

I has some time to debug this problem which today seems to happen
constinuosly and have been able to reproduce the problem with a
exim process with debuggin enabled and get this output:

.....
updating wait-remote_dk_smtp database
locking /var/spool/exim/db/wait-remote_dk_smtp.lockfile
locked /var/spool/exim/db/wait-remote_dk_smtp.lockfile
EXIM_DBOPEN(/var/spool/exim/db/wait-remote_dk_smtp)
returned from EXIM_DBOPEN
opened hints database /var/spool/exim/db/wait-remote_dk_smtp: flags=O_RDWR
dbfn_read: key=senota.com
dbfn_read: key=senota.com:1867468592
dbfn_read: key=senota.com:1867468591
dbfn_read: key=senota.com:1867468590
dbfn_read: key=senota.com:1867468589
dbfn_read: key=senota.com:1867468588
dbfn_read: key=senota.com:1867468587
dbfn_read: key=senota.com:1867468586
dbfn_read: key=senota.com:1867468585
dbfn_read: key=senota.com:1867468584
dbfn_read: key=senota.com:1867468583
dbfn_read: key=senota.com:1867468582
dbfn_read: key=senota.com:1867468581

and this continues with similar of thousands of lines like the last one,
this seems some
kind of integer overlfow error as /var/spool/exim/db/wait-remote_dk_smtp
is just 4Mb
size, maybe is due to the undelaying db libraries ?

El 12/04/2012 10:23, David Saez Padros escribió:
> Hi
>
> Today exim eat all server cpu to the point that i was unable to
> login as root. Once recovered i see one exim process eating 100%
> cpu when delivering a message to a remote host
>
> [log]# top
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> 2855 exim 20 0 42168 2412 1232 R 100 0.0 15:44.75 exim
>
> [log]# /usr/local/exim/bin/exiwhat
> 2853 delivering 1SIDML-0008Ls-OK: waiting for a remote delivery
> subprocess to finish
> 2855 delivering 1SIDML-0008Ls-OK to fltr-in1.mail.dreamhost.com
> [208.97.132.73] (kvmlcq3 [at] dub)
>
> this process never ended and i have to kill it but in the next
> queue run the same message was processed correctly
>

--
Salu-2 y hasta pronto ...

----------------------------------------------------------------
David Saez
On-Line Services 2000 S.L.
http://www.ols.es
----------------------------------------------------------------

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

exim-users at spodhuis

May 31, 2012, 8:15 PM

Post #3 of 6 (402 views)
Permalink

Re: exim delivery process hang at 100% cpu [In reply to]

On 2012-05-31 at 16:52 +0200, David Saez Padros wrote:
> I has some time to debug this problem which today seems to happen
> constinuosly and have been able to reproduce the problem with a
> exim process with debuggin enabled and get this output:

You don't mention which OS this is. Nor do you mention which version of
Exim it is.

Is it by any chance FreeBSD? If so, which release? If not, which OS is
it? What is the output of:

exim -bV

?

FreeBSD changed their DB API in r190495. Exim prior to 4.70 assumed a
part of the DB API which was once implicit. As of Exim 4.70, released
2009-11-13, Exim uses the correct API. If you're using a version of
Exim older than 4.70, then you have security issues (unless your OS
vendor is backporting fixes).

A workaround is to build with:
USE_GDBM=yes
DBMLIBS=-lgdbm
o
-Phil

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

david at ols

Jun 1, 2012, 12:12 AM

Post #4 of 6 (401 views)
Permalink

Re: exim delivery process hang at 100% cpu [In reply to]

Hi

El 01/06/2012 5:15, Phil Pennock escribió:
> On 2012-05-31 at 16:52 +0200, David Saez Padros wrote:
>> I has some time to debug this problem which today seems to happen
>> constinuosly and have been able to reproduce the problem with a
>> exim process with debuggin enabled and get this output:
>
> You don't mention which OS this is. Nor do you mention which version of
> Exim it is.

sorry, we are using Debian 6.0.3 (Linux 2.6.32-5-amd64) and Exim
version is 4.77 (self compiled)

> Is it by any chance FreeBSD? If so, which release? If not, which OS is
> it? What is the output of:
>
> exim -bV

Exim version 4.77 #9 built 10-Dec-2011 19:27:31
Copyright (c) University of Cambridge, 1995 - 2007
Berkeley DB: Berkeley DB 4.8.30: (April 9, 2010)
Support for: crypteq iconv() Expand_dlfunc OpenSSL Content_Scanning DKIM
Experimental_SPF
Lookups (built-in): lsearch wildlsearch nwildlsearch iplsearch cdb dbm
dbmnz dnsdb dsearch mysql
Authenticators: cram_md5 plaintext spa
Routers: accept dnslookup ipliteral manualroute queryprogram redirect
Transports: appendfile/maildir autoreply pipe smtp
Fixed never_users: 0
Size of off_t: 8
Configuration file is /usr/local/exim/configure

--
Best regards ...

----------------------------------------------------------------
David Saez
On-Line Services 2000 S.L.
http://www.ols.es
----------------------------------------------------------------

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

exim-users at spodhuis

Jun 1, 2012, 1:34 AM

Post #5 of 6 (398 views)
Permalink

Re: exim delivery process hang at 100% cpu [In reply to]

On 2012-06-01 at 09:12 +0200, David Saez Padros wrote:
> sorry, we are using Debian 6.0.3 (Linux 2.6.32-5-amd64) and Exim
> version is 4.77 (self compiled)

Does switching to use GDBM instead solve it?

> Berkeley DB: Berkeley DB 4.8.30: (April 9, 2010)

I'll install this and use it, see what breaks.

-Phil

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

eslist at ols

Jun 1, 2012, 3:00 AM

Post #6 of 6 (401 views)
Permalink

Re: exim delivery process hang at 100% cpu [In reply to]

Hi

by now it works well with gdbm, thanx !

El 01/06/2012 10:34, Phil Pennock escribió:
> On 2012-06-01 at 09:12 +0200, David Saez Padros wrote:
>> sorry, we are using Debian 6.0.3 (Linux 2.6.32-5-amd64) and Exim
>> version is 4.77 (self compiled)
> Does switching to use GDBM instead solve it?
>
>> Berkeley DB: Berkeley DB 4.8.30: (April 9, 2010)
> I'll install this and use it, see what breaks.
>
> -Phil
>

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads