inbound queue piling up... suddenly

[Bryan Guest]

Hello:

We have been running MailScanner for months with everything fine.  Last week 
I went on vacation...

The people who watch these things in my absence started getting warnings 
about the inbound queue(s) increasing by large amounts.

I have two Sun V20Z machines running RHEL ES4.  They each have 2GB of RAM. 
I am running MailScanner 4.56.8 and Mailwatch.

Here is what appears to happen...

There is a pile of mail in the queue that needs to be processed.
It doesn't seem to be getting processed.
I stop MailScanner.
I wait for the MailScanner processes to die off (takes several minutes).
Then I restart MailScanner.

Mail starts getting processed again.  I can see messages like this:

Mar 26 13:54:56 nicole MailScanner[7368]: New Batch: Found 1008 messages 
waiting
Mar 26 13:54:56 nicole MailScanner[7368]: New Batch: Scanning 30 messages, 
3660448 bytes

The inbound Queue starts to come down.  All appears good.

Then I hit the maximum number of MailScanner processes (65 currently).

Mail Starts to pile up in the outbound queue again.

I no longer see messages like the above, IE about new batch.  All I see are 
these messages:

Mar 26 14:11:36 nicole MailScanner[4289]: Started SQL Logging child
Mar 26 14:11:36 nicole MailScanner[5530]: Logging message l2QHUr1e026132 to 
SQL

So it would seem that MailScanner is starting processes to it's maximum and 
then stops processing mail.  And it would appear that this is because it it 
getting hung up logging to MailWatch.

Therefore I would like to ask three questions...

1.  Has anyone seen this before?
2.  Why did it just magically happen (besides the fact I went on vacation?)
3.  What can I do to correct the problem or stop SQL logging from holding up 
the show?

Many Many thanks in advance,

Bryan Guest
bryan.guest at bmts.com
bryan.guest at gmail.com