The other day my Netware 6.5.7 / Groupwise 7.0.2 server decided to stop sending out email for no apparent reason. Some of the things I tried during the troubleshooting process were:
1) Checked the GWIA log files, which didn’t show any errors occurring even with verbose logging enabled. As a matter of fact, the logs didn’t show the messages ever getting to the GWIA for processing! The MTA and POA log files did show the messages being processed, though.
2) Cleared all the GWIA queue directories, but mail still wasn’t sending out even after restarting the server.
6) I made sure nothing weird was happening with DNS lookup on the Groupwise server.
7) I went through each step in TID 10061085, ” How to troubleshoot GWIA”
8 ) As a last ditch effort, I disabled Gwava (version 3.72), which we use as an inbound spam scanner. As soon as Gwava was disabled, mail started leaving the network. I was pretty stunned, since we only scan incoming mail, and we don’t use Gwava as a virus scanner. I verified in the Gwava config outgoing mail wasn’t set to be scanned. I then re-enabled Gwava, and the mail started piling up again. I had found the culprit, but not the cause of the holdup.
I checked over the server’s Gwava log files and console screens and didn’t see any errors, but did notice a message regarding NGW-VSCAN-CONTROLLER when unloading the MTA. That led me to TID 10069173, which pointed to a corrupt message being stuck in the \domain\MSlocal\gwvscan directory. I unloaded GWIA, GWAVA, and the MTA, and renamed the \domain\mslocal directory. I restarted the server, which recreated the previously renamed directory, and mail started flowing out again.
In my case, I had a bad message stuck in the \domain\MSlocal\gwvscan\4 directory. I moved a few files at a time from the renamed directory to the new \domain\MSlocal\gwvscan\4 directory until mail stopped processing. I then downed Gwava and the MTA, deleted the problem message, then reloaded the MTA and Gwava, and mail flow returned to normal.