VMs on NFS datastores freeze when snapshots are removed

[bmekler]

I'm running 3 ESXi 4.1 update 1 hosts (essentials plus license) with shared storage on a NetApp FAS2040 accessed over NFS. PHD Virtual 5.2 is deployed for backups. Every time backup completes on a VM and the snapshot is removed, the VM in question freezes completely (doesn't respond even to pings) for 10-30 seconds. The issue is documented in the following VMware KB article:

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=103110...

However, while the article claims that the issue has been fixed in a patch, I'm still experiencing it despite running on the latest patch level (build 433742). Is there anything I should have done after installing the latest patch? If the issue isn't fixed yet, is there a workaround other than switching to iSCSI?

[rajvm256]

Hopefully you would have rebooted after installing the patch.  Can you paste the esxupdate query output?

[bmekler]

Of course, I used update manager to patch the hosts.

~ # esxupdate query

---------Bulletin ID--------- -----Installed----- -------------Summary-------------

Dell_OpenManage_ESXi410_OM650 2011-09-06T14:04:39 OpenManage 6.5 for ESXi410

ESXi410-Update01              2011-09-22T10:08:22 VMware ESXi 4.1 Complete Update 1

ESXi410-201107402-BG          2011-09-22T10:09:53 Updates VMware Tools

ESXi410-201107401-BG          2011-09-22T10:09:53 Updates Firmware

[rajvm256 · Accepted solution]

As per the KB the patch is ESXi 4.1 patch ESXi410-201104402-BG contains vmware-tools. One thing to try might be to upgrade your vmware-tools in the VM? (I am not sure if you have upgraded it)

[bmekler]

That was it! I checked the VMware tools version and it reported as 8.3.7 build 341836. Despite the vCenter reporting vmware tools status as "okay" rather than "out of date", when I ran automatic vmtools update, the version number changed to 8.3.7 build 433358; when the backup ran, I monitored the upgraded VM with ping -t and it only dropped a single packet when removing snapshot.