I have 1 cluster of 4 servers (dell 1950's) running 3.0.2 and a NFS central VOL. We run our own script which quieces the VM's and SNAPs them via NetApp. We disabled the script last week because it was generating delta files for the VM thus causing rapid disk growth. Sometimes this happens and we just roll the deltas back in and reboot the offending guest (or host). Sometimes just restarting the mgmt and VPXA services fixes the issue.
I have deleted all the Delta files and some zdump files that were present. I also took the opportunity to clean up the VOL from a few unregistered VMs and empty folders from other admin's in the system.
Issue is I am still seeing disk growth of 10 GB a day min.
Question 1... once you allocate a disk to a guest does allocate the entire amount so you will not see growth ? Or will it grow as the disk fills up ?
I am wondering if this is what I am seeing.
Are you looking at the growth in space from the netapp side?
In VMware thickly provisioned disks pre-allocate at creation time, but do not write these blocks down to disk until required. Thus from the netapp side the blocks will not show as utilised until they are written.
This is a pretty good explanation of thin provisioning disks: http://blogs.netapp.com/virtualstorageguy/2009/10/vce-101-thin-provisioning-part-1-the-basics.html#m...
Your snapshot script should quiesce the VM by creating a VM snapshot, create the netapp snap, remove the VM snapshot. I have a script here doing exactly that and it's working very well.
Are you rotating the netapp snapshots?