We have ESX 3.0.2 and VC 2.5 with NFS NetApp VOLS for storage. We run a script to quiesce, snap (NetApp not VM snap), then write back. As of late this has been failing at the write back stage. This creates a Delta file every time it fails and then we have to write them back manually. However 1 Guest escaped us and its now at TB in size. We are working on folding it all back is when the weekend comes.
My boss who likes Xen better (and has the other side of our team working on it) wants to know why we cant just snap and skip the quiesce if were not concerned about those few seconds of data we would miss. Our SLA is only to provide a guest backup from the previous day.
I think that if a file is being written to it will not finish the commit and cause corruption rendering our backup useless.
Any thoughts appreciated
SG
<!-- Kerio Popup Killer - script has been appended by KPF -->
<!--
nopopups();
//-->
<!-- Kerio Popup Killer - end of the script appended by KPF-->