Urgent issue: webmail/webadmin randomly not working!

[nosebreaker]

It's the strangest problem, I can duplicate it on the server itself by going to http:// localhost:8080 or https:// localhost:7071, I see that it is listening on 8080 and 7071, but when I try to hit it via browser, sometimes it works and sometimes not! The box isn't doing much, it just times out. A reboot doesn't do anything.

I'm not sure if this is related or not, when I try from the outside it just doesn't work, then I logon to the server and open a browser, then it usually works. Then I try from the outside again and that works suddenly too. For a minute or two anyway. It also times out on the server again.

I don't see anything in zimbra.log or /var/log/messages, any ideas?

[y@w]

Have a look in /opt/zimbra/log/mailbox.log for any weird things happening.

[nosebreaker]

I don't see anything spurious here, any suggestions as to what to look for? I see lots of soap requests from me and another user, oh wait it just did it again, i see:

2010-09-02 00:00:03,289 INFO  [btpool0-10://www.mydomain.com:8080/service/soap/SearchRequest] [name=myuser@mydomain.com;mid=7;ip=173.166.71.41;ua=ZimbraWebClient - IE8 (Win)/6.0.7_GA_2473.RHEL5;] soap - SearchRequest
2010-09-02 00:00:08,284 INFO  [btpool0-10://localhost:7071/service/admin/soap/AuthRequest] [ip=127.0.0.1;] soap - AuthRequest
2010-09-02 00:00:08,428 INFO  [btpool0-10://localhost:7071/service/admin/soap/GetAllServersRequest] [name=zimbra;ip=127.0.0.1;] soap - GetAllServersRequest
2010-09-02 00:00:25,519 INFO  [MailboxPurge] [name=myuser@mydomain.com;mid=7;] sqltrace - Slow execution (2301ms): DELETE FROM mboxgroup7.open_conversation WHERE mailbox_id = 7 AND conv_id IN (  SELECT id FROM mboxgroup7.mail_item AS mi  WHERE mi.id = open_conversation.conv_id AND mi.mailbox_id = open_conversation.mailbox_id  AND date < 1280721623)
2010-09-02 00:00:48,065 INFO  [Timer-Zimbra] [] SessionCache - Removed 2 idle sessions (SOAP, ADMIN). 3 active sessions remain.
2010-09-02 00:00:57,065 INFO  [btpool0-12://mail.mydomain.com:8080/service/soap/SearchRequest] [name=anotheruser@mydomain.com;mid=10;ip=209.6.76.155;ua=ZimbraWebClient - FF3.0 (Mac)/6.0.7_GA_2473.RHEL5;] soap - SearchRequest

[nosebreaker]

Ok, I just tried to access the backup server, and I was having the same problem. So I connected with a VPN to another office, and tried from there, and it worked fine. So something else is going on here I think.

Although now the primary claims antivirus isn't running, and the backup claims stats isn't running. Any idea how I can try to start those? I believe antivirus is running, since it was sending me messages about how it blocked virus emails earlier today.

[nosebreaker]

Disregard, its happening again. I can't even get the login page on the server itself, yet it says it is listening on those ports.

[nosebreaker]

It's driving me mad, it works, then it doesn't. I need to take a break from this, will check later. At least its accepting mail!

[nosebreaker]

Alright, its still randomly not working, anybody have any ideas?

[Hivos]

A double ip-address in your network would cause the exact same issue, so perhaps it's a network problem? You could try to access it directly from your server via e.g. lynx / elinks and see if you still have the same problem.

[nosebreaker]

There is only the one server on the network, it is the only device behind the firewall! I even replaced the firewall with another device with no luck. Both the primary AND backup server were doing the same thing! Different hardware, slightly different setup, not sure why it was doing it.

At the time, I couldn't access it via its local IP either. It was just as flaky trying to talk to itself. I could SSH in fine, hit apache, etc.

For some reason, the problem is resolved on both boxes now! Now and again I have to refresh a page, but it works 95+% of the time.

[nosebreaker]

So the problem has been coming back and going away. It was never solved, does anybody have any ideas on this? I'm about to give up and go back to postfix/squirrelmail!