Internet Mail gateway not processing messages - Take 2

[easysoft]

Hi All,

This problem has reoccurred after barely a week .... my previous posting was here :

viewtopic.php?t=8581

Again, I have messages stuck in the UNIX queue, but this time I can't just zap the messages in the queue. Also, this time there are no missing messages.

I have tried shutting down all services I can from the SAC. and running :

# omscan -A -Q UNIX -v -f

But I always get the error :
omscan : [OM 4951]
A serious error has occurred. Please see the log files.
Event logged on 31.08.07 at 10:47:21.
Owner/Context Info : UNIX
Additional error info:
omscan : [OM 11170] Could not obtain exclusive access to a queue.

Any ideas ? I'm quite stumped now. Something must be wrong in the UNIX queue but I have no idea how to fix it ....

Help !

Thanks in advance,
Arif

[mikethebike]

Arif,

omoff -d 0 -s unix

ps -ef |grep -i unix

if there are any unix.out processes, kill them (try kill <pid> first, if that does not work, use "kill -9").

Use omqdump to move all the messages to the error queue (make sure you clear out the error queue first (use the "M" option to move multiple messages).

Start the unix service (omon -s unix).

Check how many queue runners you have :

omstat -s -p

you may wish to increase the number of queue runners to make the queue run better? If you want to do that, stop the service, then :

omsetsvc -x unix 3
(creates 4 queue runners).

Now use omqdump to move messages back in batches of, lets say 200, to the unix queue (using the "M" option again.

Keep an eye on the queue, and the fatal/event logs for any errors.

You may even want to remove some messages (omstat -q error) that are obviously spam etc.

Mick

[easysoft]

Thank you very much Mike !

I tried what you suggested - and the omscan now runs succesfully on the queue.
I have started moving mails back from ERROR to UNIX and they are passing through.

Oddly though, I'm sure they started moving out even before I ran the omscan - so what could have been the problem ? Could a simple shuffling between queues have been enough to jolt things into action ?

Again, many thanks,

Arif

[mikethebike]

Arif,

when the service is started, it looks through the queue to decide which messages should be processed first (time, date, priority etc). If the queue gets very big, it takes so long to check through the queue, it appears to hang,
Increasing the number of queue runners helps, and is worth looking at. Remember though, that increasing the number of queue runners may impact perormance of other services, so its a bit of a balancing act.

So...the bigger the queue, normally the slower it will run.

Mick