[SOLVED] Queue Temporarily Suspended - can't connect to 127.0.0.1

[dwmtractor]

This situation may be related to this old thread (which doesn't really show a solution), but since I saw it today with 4.5.9 I wanted to post this for comments and info.

This morning I had my entire MTA queue suspend everything that was supposed to be delivered. Each message had a status in the mail queue of "Delivery temporarily suspended", and here's a sample entry from /var/log/mail.log:

Nov 21 10:58:14 mail postfix/qmgr[1288]: 4DA9B3D0C57: to=<someone@mydomain.com>
, relay=none, delay=3645, status=deferred (delivery temporarily suspended: connect to
 127.0.0.1[127.0.0.1]: Connection refused)

zmcontrol stop and start did not accomplish anything, and neither did zmfixperms. However, the thread above suggested "something to do with amavisd", so I tried disabling, and immediately re-enabling, the "Enable spam checking and antivirus" checkbox from the AS/AV tab, global settings, in the UI. Within a minute of doing this all the deferred mail went through.

I don't have any idea what caused this to hang; however, earlier this morning I changed my spam settings to a lower tag percent (using the UI), and did NOT stop/restart spam filtering. It appears that the queue hung when I did this.

There may be a bug to report in this; I'll wait for further confirmation from other users first. But in the meantime, if you find your formerly-working MTA suspended, try stopping and restarting antispam/antivirus from the UI and see if it kicks it loose for you.

Happy Thanksgiving everyone!

Dan

Note: Please see this wiki for complete instructions and descriptions of this issue. The problem may also manifest itself with the warning:

WARNING: Your ClamAV installation is OUTDATED!
WARNING: Local version: X.XX Recommended version: Y.YY
DON'T PANIC! Read `http://www.clamav.net/support/faq`

[jbareMACS]

We experienced the same problem overnight and in to this morning, with the same errors in /var/log/mail.log. The only conclusion I came to was that it was related to the spam/AV services.

On Thursday evening I upgraded the server to 4.5.10. On Friday evening, I realized that the spam and anti-virus services were running (we have ProofPoint as a separate perimeter MTA, and it handles spam & AV). I unchecked the options for "Anti-Spam" and "Anti-Virus" on the Services tab of the Server admin web interface and clicked Save. However, I never did a stop/start of zmcontrol.

A co-worker rebooted the server today, which seems to have resolved the issue.

This sounds like a bug with reproducible results, based on dwmtractor's post and the issue we experienced today.

[jholder]

ClamAV can be a bit picky. Sometimes it will download a virus update, and the MD5 will be wrong. If it is, then ClamAV usually goes down, and the system suspends delivery.

You'll know this is the case if you go to the /var/log/zimbra.log and you see that ClamAV isn't running.

To fix this, you can delete the defs, and try again:

  su - zimbra
  mkdir /tmp/clamdb
  mv /opt/zimbra/clamav/db/* /tmp/clamdb
  zmprov ms `zmhostname` +zimbraServiceEnabled antivirus
 /opt/zimbra/clamav/bin/freshclam --config-file=/opt/zimbra/conf/freshclam.conf 
 zmamavisctl stop

zmamavisctl start