local delivery 5k+ and not moving

[RickC]

Hi,

I have a problem with local delivery on one of my back end servers. I cannot get the local delivery queue to empty, all services are running and there are no errors in the logs.

I did have an issue with Clamd earlier. Clam caused smtp to abort.

ERROR Service Router(Service Router) Thu Oct 27 15:02:19 2011
[OM 5181] Reply timed out or invalid - Mapper protocol problem.
Command sent: <none - expect greeting reply>
Reply received: 503 "ClamAV" is not running
Pid of logging process: 12284

I had to disable it :

1039 cd $(omrealpath "~/rules")
1040 pwd
1041 ll
1042 mv ALL-ROUTES.VIR ALL-ROUTES.VIR.not
1043 omoff -w -d0 sr
1044 omon sr

Is it possible Clamd is also causing issues in the local? smpt is empty now.

Thanks! Running 11.4.4 on RHEL4

[RickC]

An update, the local delivery queue is going down VERY slowly. Over 3 hours we've gone down 1,000 messages to approx 4,000.

There is absolutely no activity on the server, top shows java as being the main app running:

top - 20:11:39 up 3:44, 2 users, load average: 9.50, 9.09, 8.36
Tasks: 371 total, 4 running, 367 sleeping, 0 stopped, 0 zombie
Cpu(s): 17.5% us, 27.4% sy, 1.2% ni, 29.9% id, 23.5% wa, 0.0% hi, 0.6% si
Mem: 16437420k total, 16383476k used, 53944k free, 1006524k buffers
Swap: 18448376k total, 0k used, 18448376k free, 10881848k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
7840 root 18 0 2566m 2.1g 16m S 65.4 13.5 63:05.57 java
5717 scalix 16 0 20392 5576 4060 R 27.2 0.0 1:45.27 mime.browse
2478 scalix 17 2 45612 29m 6632 D 24.6 0.2 41:41.70 local.delivery
2479 scalix 17 2 45216 29m 6044 D 24.2 0.2 42:00.28 local.delivery
2480 scalix 17 2 45964 29m 5688 R 23.6 0.2 43:04.89 local.delivery

Rick

[RickC]

Queues cleared last night, everything back to normal.

Though wish I knew what the cause was, wasted 1/2 a business day with the mail stuck.

[mikethebike]

Hi Rick,

when things are bad like that, I would normally move ALL the messages from local queue to something like smerr or error queue ~(not the TEST queue), then move them back in in say 200 or 300 batrches using omqdump. The queue processes a lot quicker with fewer items.
Other thing to check is number of queue runners you have configured, that can make a big difference...I usually ran with 4 or 5 aux runners.

Mick

[RickC]

Thanks Mike, I'll keep that in mind next time!