Disk usage goes from 99% back to 13%

[rwcanary]

I'm having a wierd issue happening here. The disk usage will peak out at 98% and then drop back down to 13%. During that time zimbra because unstable and different errors happen during that time.

If shut down zimbra and do a "ps aux | egep zimbra" there are about 500 zimbra users trying to perform a df. Which I assume is going into a loop because system resource have been depleted. After it sets awhile it will finally clears out and the disk usage goes back down to 13%.

I believe this is a mass email surge, but I am unsure how track its origin. Any suggestions?

Please see attached mailbox.log

[veronica]

Can you attach mailbox.log ?

[rwcanary]

trying to.... uploaded the log file, but I seemed to forgotten how to get it attach

[rwcanary]

I attached it to the original post. I had to compress it, thats why it wasn't taking it the first time.

Thanks

[veronica]

These log seem to me when ldap service has alredy gone down:-

2010-03-24 00:00:48,409 WARN [threads.csv] [] perf - Unable to determine thread name prefixes.
com.zimbra.common.service.ServiceException: system failure: ZimbraLdapContext


Can you post snippet of before this happened to figure out what caused ldap to go down