Openmail not delivering mail

[James Meadows]

We use OpenMail B.05.10 on HP-UX 10.10. Today it decided to no longer deliver mail to our users. Outgoing mail seems to work. Incoming mail seems to go to the ROUTER queue (we see over 1000 messages if we do omstat -q router). These appear to be the messages that aren't getting delivered. The mail isn't getting bounced back, and the server doesn't report any errors back to the sender. We can use sendmail to send messages to local HP-UX accounts, and the POP3 interface seems to function, but nothing ever appears in an Openmail user's "intray." We're not even sure how to figure out where these messages are or why they're not being delivered. Any help would be appreciated. And please, reply to jrmead@citynet.net, since that's the only address I have that works right now!

[Robin Wakefield]

Hi James,

Are the messages stuck in the ROUTER queue? If so, you may have a bad message at the top. Or are they getting dumped into an error queue for some reason. If so, use omqdump to see. Could they be getting dumped into xport.hold?

Is there anything in the fatal log? Do you see anything with omshowlog (use a high logging level if necessary).

Has anything changed to possibly cause this??

Rgds, Robin.

[James Meadows]

Robin,

Yes, the messages are stuck in the Router queue. If I do an omstat -s it says that I have over 2000 messages in the queue. An omstat -q "router" will list all the messages but the first 10 lines have the following message:

omstat : [SYS 2] No such file or directory

The mail logs and syslogs look okay but maybe I've overlooked a log somewhere?

Everything worked fine until about 10:20 a.m. yesterday, when mail stopped being delivered locally. I can send mail fine and it will deliver it to other external servers.

There is another note of possible interest here. I ran fsck yesterday evening and it found and fixed quite a few problems. I didn't think too much of it because I wasn't sure how long it had been since my recently departed SysAdmin had done this. However, the server decided to reboot itself last night and entered single-user mode on reboot. We ran fsck again and found numerous additional errors. Since then, we have ran fsck several other times. We also run it twice in a row every time. It never finds more errors on the second run, but may find some an hour later. Could the Service Router hangup be related to a file system or disk problem?

James

[James Meadows]

I spoke with HP Support today and did more troubleshooting. I don't know what caused the initial problem, but it seems that some message stuck in the queue and things started backing up. I soon had over 3000 messages in the Service Router queue and my filesystem filled up.

I stopped the Service Router service and moved some messages to the error queue and restarted it. This got my messages started but they were going amazingly slow. I finally moved 3000 messages to the error queue and have spent the evening moving them back in batches of about 300. This doesn't seem to bog things down too much and it doesn't put our current email too far behind in its delivery.

If nothing else, this experience has introduced me to many commands that begin with 'om'.

James