Admin login stuck on "Loading..."

[danfluidmind]

I just did a fresh install of Ubuntu 6.06 into a virtual machine and installed Zimbra 4.5.5. It installed fine and started up fine. But when I connect to port 7071 with a web browser and try to login to my admin account, the browser just hangs on "Loading...".

Anyone ever have this problem and know what's causing it?

I tried putting an incorrect password to see if it came back with a "login failed" message, and it did. It only hangs when the login was successful. The httpd_access.log file is completely empty. Here's what's in the audit.log (admin account renamed):


2007-06-28 19:53:02,652 INFO [http-7071-Processor48] [ua=ZimbraWebClient - FF2.0 (Mac)/undefined;ip=74.138.218.234;] security - cmd=AdminAuth; account=admin@example.org;
2007-06-28 19:53:02,662 WARN [http-7071-Processor48] [ua=ZimbraWebClient - FF2.0 (Mac)/undefined;ip=74.138.218.234;] security - cmd=Auth; account=admin@example.org; protocol=soap; error=authentication failed for admin@example.org;
2007-06-28 19:53:02,662 WARN [http-7071-Processor48] [ua=ZimbraWebClient - FF2.0 (Mac)/undefined;ip=74.138.218.234;] security - cmd=AdminAuth; account=admin@example.org; error=authentication failed for admin@example.org;
2007-06-28 19:53:05,286 INFO [http-7071-Processor47] [ua=ZimbraWebClient - FF2.0 (Mac)/undefined;ip=74.138.218.234;] security - cmd=AdminAuth; account=admin@example.org;
2007-06-28 19:53:05,289 INFO [http-7071-Processor47] [ua=ZimbraWebClient - FF2.0 (Mac)/undefined;ip=74.138.218.234;] security - cmd=Auth; account=admin@example.org; protocol=soap;

Thanks
--Dan

[jholder]

Hi, and welcome to the forums!

All that tells us is that you entered the wrong password.

Is this FireFox? If so, can you look at the error console to see if there are any errors?

[danfluidmind]

Originally Posted by jholder

All that tells us is that you entered the wrong password.

Thanks J. :-)

I guess I didn't make that clear. Every time I try to login with the wrong password, it does just as expected: comes back to the login screen and gives an "authentication failed" message. It does that every time, so there's no problem with trying to login with the wrong password. The first three log entries above were from one of those attempts.

When I enter the correct password is when it freezes on "Loading...". The last two entries in the log entries above were from one of these attempts to login.


So last night I left a browser in that "Loading..." state. This morning, the VM that Zimbra is running on was EXTREMELY sluggish (hit return at a command prompt and it takes 30 seconds to give you anothre prompt). I brought up the VMware console and on the screen for that VM is has the following messages:

Out of Memory: Killed process 6981 (java).
Out of Memory: Killed process 6982 (java).
Out of Memory: Killed process 4118 (java).
Out of Memory: Killed process 15381 (java).
Out of Memory: Killed process 3977 (mysqld).
Out of Memory: Killed process 3977 (mysqld).
Out of Memory: Killed process 15505 (java).
Out of Memory: Killed process 15545 (java).
Out of Memory: Killed process 15532 (java).
Out of Memory: Killed process 15534 (grep).
Out of Memory: Killed process 15534 (grep).
Out of Memory: Killed process 15678 (java).
Out of Memory: Killed process 15726 (java).

So I closed the browser and those messages stopped and after a few minutes it started responding enough to get "top" running. The gig of swap space is 100% used. The CPU is 0.0% idle (6.9% sy, 92.1% wa). The java process keeps showing up on top.

Something's gone haywire!

[danfluidmind]

Originally Posted by danfluidmind

So I closed the browser and those messages stopped

I take that back. I'm still getting "Out of Memory: Killed process ##### (java)." every few minutes and it's still extremely sluggish. I just entered the command to shutdown Zimbra. Waiting for it to actually do it :-)

[phoenix]

It would seem the obvious question but I'll ask it anyway - how much memory is allocated to this VM?

[danfluidmind]

Originally Posted by phoenix

It would seem the obvious question but I'll ask it anyway - how much memory is allocated to this VM?

1 gig.

This was just installed yesterday and has no users on it yet. Obviously I haven't been able to login to the Admin site yet. The /opt/zimbra directory is an NFS share being exported from the VM host OS through vmnet1.

--Dan

[danfluidmind]

Okay, the Zimbra server processes finally finished shutting down and RAM and CPU usage is back to normal so I can finally look around at the command prompt.

Looking through /var/log/zimbra.log (I didn't realize that was there), I see a bunch of messages like these throughout the night:

postfix/showq[31035]: fatal: select lock: Resource temporarily unavailable
postfix/tlsmgr[31034]: fatal: cannot lock PRNG exchange file /opt/zimbra/postfix-2.2.9/conf/prng_exch: Resource temporarily unavailable
Jun 29 01:30:24 plato postfix/master[4252]: warning: process /opt/zimbra/postfix-2.2.9/libexec/showq pid 31035 exit status 1
postfix/master[4252]: warning: /opt/zimbra/postfix-2.2.9/libexec/showq: bad command startup -- throttling
postfix/master[4252]: warning: process /opt/zimbra/postfix-2.2.9/libexec/tlsmgr pid 31034 exit status 1
postfix/master[4252]: warning: /opt/zimbra/postfix-2.2.9/libexec/tlsmgr: bad command startup -- throttling
slapd[3753]: is_entry_objectclass("", "2.5.6.1") no objectClass attribute

These messages occur over and over again throughout the night (like once a minute).

Where else should I look that might help indicate what was going on?

Thanks
--Dan

[jholder]

Can you

cat /dev/null > /var/log/zimbra.log

Then do a fresh startup, and send me the log?
jholder@zimbra.com

Looks like some sort of problem with TLS

[danfluidmind]

Thanks for trying to figure that out John.

I went ahead and did a complete re-install of Ubuntu into a new virtual machine, and this time installed Zimbra on the virtual disk and it's working fine. It seems to me that it had to do with the fact that I installed Zimbra onto an NFS share mounted at /opt/zimbra

So, for future reference...don't do that :-)

--Dan

[jholder]

Yep-
That'll do it. We don't support NFS mounting