Internet Mail Gateway segfault

[jtrott]

We have the Internet Mail Gateway segfaulting every time we start it up.
From /var/opt/scalix/logs/fatal:

SERIOUS ERROR Internet Mail (Outgoing ) Tue Dec 12 21:27:59 2006
[OM 10270] Process about to terminate due to error.
Signal (Segmentation Violation) trapped by process 9583
Procedure trace follows:
-> tf_OpenFile
-> im_ItemRef2FName
<- im_ItemRef2FName
-> tf_FindFreeEntry
<- tf_FindFreeEntry
<- tf_OpenFile
<- Msg_GetTfHandle
-> tf_FlushFile
<- tf_FlushFile
-> Msg_GetRef
<- Msg_GetRef
-> mpl_AddMsgToQueue
-> gen_GetLock
<- gen_GetLock
-> mpl_OpenPoolFile
<- mpl_OpenPoolFile
Pid of logging process: 9583


SERIOUS ERROR Internet Mail (Outgoing ) Tue Dec 12 21:27:59 2006
[OM 10272] BACKTRACE:
/opt/scalix/lib/libom_er.so(er_add_backtrace+0xc6)[0xdc3f06]
/opt/scalix/lib/libom_er.so[0xdc41d6]
/opt/scalix/lib/libom_er.so(er_DumpProcAndExit+0x1f)[0xdc437f]
/lib/tls/libpthread.so.0[0x129898]
/opt/scalix/lib/libom_q.so(Q_PutMsgOnQueue2+0x187)[0x221123]
unix.out[0x804b74a]
unix.out[0x804b0ec]
unix.out[0x804ae9b]
/lib/tls/libc.so.6(__libc_start_main+0xd3)[0x8adde3]
unix.out[0x804a1c5]
Pid of logging process: 9583

The output of omstat -s is:

Service Router Started 20:46:27 0
Local Delivery Started 20:46:27 0
Internet Mail Gateway Started 21:44:53 945
Local Client Interface Enabled 20:46:27 0
Remote Client Interface Enabled 20:46:27 2
Test Server Started 20:46:27 0
Request Server Started 20:46:27 0
Print Server Started 20:46:27 0
Bulletin Board Server Started 20:46:27 0
Background Search Service Started 20:46:27 0
CDA Server Started 20:46:27 0
POP3 interface Started 20:46:27 0
Omscan Server Started 20:46:27 0
Archiver Started 20:46:27 0

The error queue has over 6000 items in it.
I've rerun the sendmail configuration, made sure sendmail was the preferred MTA (CentOS 4.4), and enabled sendmail debugging in the unix.mapper file (hasn't been modified since the problem started).
Any ideas on what next to try and get outgoing email working again?
Thanks,
JT

[ScalixSupport]

Hi!

Have a look at these links hope they help:

http://www.scalix.com/community/viewtop ... l+outgoing
http://www.scalix.com/community/viewtopic.php?t=3350

NB: BTW what version of Scalix you are using.

Thanks.

Warm Regards,
Subir

[jtrott]

Nope, I'd already searched the forums dry before posting.
Version is 10.0.1.3.

Here was the solution and the problem.

When configuring Scalix for the first time on the site I had enabled sendmail debug logging with the transport-service = /usr/sbin/sendmail -bs -X/tmp/scalix-sendmail.log line in the unix.mapper file. I'd not thought about it after that point as it was unrelated to the problem I was troubleshooting at the time.
What happened was the log filled up until it was 2GB in size. The filesystem, ext3, decided that it couldn't get any bigger and sendmail stopped working. At that point I think Scalix started to get corrupt messages in the outbound mail queue due to sendmail barfing on the log file and the end result was no email out of the system.
As soon as we cleared the queue and deleted the log file the system started working again.
I guess the lesson here is not to leave the debugging logging enabled once a problem has been solved.
Thanks,
JT

[ScalixSupport]

Hey JT,

A good catch, the maximum size for a log file is 2GB. Great that things have started working again.

Warm Regards,
Subir