ConfigServer Community Forum

I realize this forum isn't an official avenue of support, but I thought I'd start here, first, and see if anyone else is having this problem. After upgrading to MailScanner v2.85 (just came out today), MailScanner will not start/stay started, even though I keep getting an "OK" after attempting to start/restart from SSH.

I've been through all the technical FAQs to troubleshoot:
http://www.configserver.com/techfaq/index.php?faqid=66
http://www.configserver.com/techfaq/index.php?faqid=14

I've run: /usr/mscpanel/mscheck.pl numerous times. I've rebuilt exim, reset it to defaults, restarted exim, MailScanner, etc. Forced a reinstall of MailScanner, twice. Nothing works. I've never had this problem with any other version of MailScanner, and I've used it on several servers for years. I finally disabled it, so as not to hold up email delivery, but obviously I want it fixed rather than simply disabling it.

Is anyone else having an issue with this version?

The only possible unresolved error I can see is "trying to setlogsock (unix)."
When I tail the /var/log/maillog I see "MailScanner[28308]: Using locktype = posix" where I previously used to see it was using "flock." I commented out the "lock_method flock" in both spam.assassin.prefs.conf and local.cf, hoping this would fix the issue, but doing that hasn't fixed it (and/or has "buggered" it up further.

The previous version of MailScanner was working fine. Is there a way to roll-back to that version?

Updated 2 boxes with CentOS 5.8 and didn't have an issue, which of course doesn't help you.

Since you've run through all the obvious options you might submit a ticket although presently it's after 11 PM where ConfigServer is located.

You could try the MS mailing list as there are some very knowledgeable and helpful individuals who frequent the list.
http://lists.mailscanner.info/mailman/l ... ailscanner

Good luck. Wish I had more to offer. Would be interested to know what the solution turns out to be.

@MaraBlue,
try to restart MailScanner and do a tail -f at /var/log/maillog file.

Post all the info that MailScanner shows in maillog at startup.

That could help to check what is wrong.

Thank you sawbuck and Sergio, I really appreciate both your responses. I was able to get it started again last night, though what I did shouldn't have made any difference. The initial error I got was that MailScanner was starting yet the MailScanner child was dying of old age after 1 minute of running. MS *would* start up if started from the command line (and then die), yet not start at all if started from MSFE. Each time from the command line I'd get a "started [OK]" message, but it would die again.

I went through the basic troubleshooting steps (stopping then restarting MS, running /usr/mscpanel/mscheck.pl, rebuilding the exim config, checked the perl modules, running lint on SA, etc), and each step seemed to introduce a new error that hadn't previously been a problem (new Exim config error, Cannot create + lock clean tempfile, etc). We migrated to a new server 1.5 weeks ago and reinstalled everything, but it was all working fine on the new server and on the previous server for years.

I can't rule out there wasn't an issue with the Exim config file before I rebuilt it (using /scripts/buildeximconf) and yet it was working fine before yesterday, and I know for a fact Exim had been restarted within the last week.

Following basic troubleshooting procedure, I tried one thing at a time to isolate the problem, correctly each error that came up. The very last thing I tried (before I was going to uninstall/remove MailScanner then reinstall with a forced checkperlmodules) was to comment out the "lock_method = flock" in the /usr/mailscanner/etc/spam.assassin.prefs.conf and fixed/set the path to DCC. AFAIK, neither one of those should have caused MS to fail to load. I tried starting MS once more and it stayed running. I guess the 15th time is the charm?

It's a mystery, and will perhaps always be one, but if no one else is having issues with the latest MS update, then it must have been something I did. Maybe some odd network connection issue that caused a corrupt file during the upgrade? Solar flares? Sunspots?

Anyway, thanks so much for the replies.

Glad you got it working Mara.

Have often had concerns about updating MS (which is why we try to do it during non business hours) mostly because cPanel's Exim implementation often has unintended consequences.

ConfigServer has done and continues to do an excellent job keeping MS in sync with cPanel.

sawbuck wrote:Glad you got it working Mara.

Have often had concerns about updating MS (which is why we try to do it during non business hours) mostly because cPanel's Exim implementation often has unintended consequences.

ConfigServer has done and continues to do an excellent job keeping MS in sync with cPanel.

I totally agree on all counts.