Yes, and even after restart, the problem seems to quickly re-appear.
Originally Posted by Vladimir
The remedy that lasts the longest is a complete re-install/upgrade of the same version that's running. Since I'm hosting this VM in the Amazon cloud, it was fairly painless to update the machine to a more powerful VM. Strangely, the problem still happens in about the same amount of time.
I've read some anecdotal reports of similar issues caused by the size of the disk write buffers in the VM coupled with the somewhat slow speed of the Amazon EBS volumes. If the buffer size is based on a percentage of total RAM, larger instances try to flush much larger chunks of data less often. I'm no expert on this stuff, but perhaps someone on this list has had similar experiences?
For what it's worth; I installed Zimbra 7.2 on a CentOS 6.3 VM in our vSphere environment, migrated the data from the old Zimbra 6 machine to it using ZeXtras Backup Suite, and everything is running great. We have about 50 users on the VM with 1 vCPU and 4GB RAM, Avg. Load is well below 1 at all times, CPU usage is below 25% most of the time.
Good to know. I'm thinking that by bumping up my VM size, the increased Memory results in larger IO buffer size, and I'm actually seeing the problem occur more often than when it was on the smaller instance...