Deferred Queue shows messages, scan produces no result

[rkerr]

Hi Forums

using the admin ui I can get results off of the active queue and other queues but the deferred queue never returns from scanning. I have recently upgraded from 6.0.10 to 6.0.13.

Following
Mail Queue Monitoring - Zimbra :: Wiki

I have done fixperms and have reset the ssh keys.

when listing the queues, it shows counters correctly.

What next?

[phoenix]

Originally Posted by rkerr

using the admin ui I can get results off of the active queue and other queues but the deferred queue never returns from scanning. I have recently upgraded from 6.0.10 to 6.0.13.

I'm not really sure what you're saying here. Do you actually have mail in the deferred queue? If you have then go to the log files and see what the problem is.

[rkerr]

I had mail in the deferred queue.
If I clicked on 'Mail Queues' in the admin UI it would show me '55' in the deferred queue.
When I would follow that up with a click on 'myserver.example.com' the 'Deferred' Tab would show 55 but the screen would never fill. Eventually I would get a 'EMPTY_RESPONSE' message.
Nothing in the log files about 'EMPTY_RESPONSE'.
The funny thing is that I checked the active queue and it returned 2 results.

Even better, I checked an hour later and it was working again. On that alone we could probably close the thread.

This server has been incredibly unstable since our recent changes.
1. Upgraded 6.0.10 ubuntu 8 x64 to 6.0.13 ubuntu 8 x 64
2. moved zimbra out of /etc/init.d
3. did an operating system upgrade from Ubuntu 8 x64 to ubuntu 10 x64
4. reinstalled 6.0.13 ubuntu 10 x 64
5. reboot
(Note: zimbra is running as only VM on a ESX server, had to do it this way as Ubuntu10 was not release at the time of setup and my disk controller was having huge issues with ubuntu 8)

So far it has been nasty to get going. 3 hours fighting with bind and not having it load my zone file. ripping out bind and going with dnsmasq. more time reissuing ssh keys. even more time adding root ulimits that were somehow missed by the installer. Big performance issues with amavis only remedied by adding a ramdisk. Permissions getting corrupted requiring a zmfixperms. I've been using this product on multiple machines since 4.05 and have never has so many issues. Yes I'm venting here Performance is much improved over previous OS/version combo so I'm hopeful on getting it stable.

The issue originally posted showed up after getting some auth issues with the queues. I reissued the keys. I guess the admin interface took its time to start using them for the deferred queue or something.

Crossing my fingers and hoping something new doesn't show today. What would be great at this time would be some knowledge on how to force a full log rotation without having to get my hands too dirty so that I can start with logs not contaminated with previous errors to sift through. Documentation I have found on logs seems to indicate that forcing the system to start a fresh set of logs would not be trivial.

Any suggestions on other places to look for issues? System finally seems stable but more users will start coming on in another hour or so.