VMware ESX has a great feature that pauses VMs when the VM's underlying storage runs out of space (or possibly is dangerously close to doing so). Is there any way for ProxMox to do the same thing?
I just got around to testing this. I setup a VM with werror=enospc added to the scsi0 string in the config file. I created a qcow2 disk with the disk size > the host volume size. I then ran dd if=/dev/urandom of=test.img status=progress and waited and watched. I also watched the host to see how it was doing. I tested on both NFS and GLUSTERFS storage backends. When the host storage ran out of space, the following happened:
On NFS - VM WAS AUTO-PAUSED - Not reflected in UI as being paused. Only option is "Start", which fails after doing nothing. Needed to unpause using qm resume from the command-line.
On GLUSTERFS - VM CRASHED - Reflected in UI - Needed to restart.
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.