We recently enable kRBD (tick the KRBD box in storage configuration) for our running cluster. Faster IO was the main reason, We record 30% gain in reads/writes in all VM running in writeback cache mode.
But then a couple of VMs disks needed to be resized. In both cases disks went to 0 bytes...
It might be because of my setup (SAN / iSCSI) but I can only use IDE on Windows VMS and BUS SCSI with "VirtIO SCSI" controller on the Linux ones. Any other combination make the disk not available during OS installation.
About the VMs shutting down, I have to wait until the weekend to say if...
Hey Guys, I'm having the same issue once a week, and it only affects Windows VMs.
I'll try unchecking 'discard' and changing to "write back" as a test in one of the VMs and report back.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.