gpordeus commented on PR #7615: URL: https://github.com/apache/cloudstack/pull/7615#issuecomment-1693829103
> @gpordeus I can reproduce the issue @vladimirpetrov faced in his testing > > image of vm B does not have backing file either, but it works (the disk size is much bigger than vm A) @weizhouapache Great, thank you for testing and for pointing out the disk size. I hadn't been able to replicate it before [last week](https://github.com/apache/cloudstack/pull/7615#issuecomment-1684389067) because I was testing from and to iSCSI and NFS storages, while this error only happens in NFS to NFS. > @DaanHoogland @rohityadavcloud @gpordeus > > my summary: if template is removed, > > Before this PR, > > * when migrate volume of stopped vm, it succeeds. vm also works after starting > > * when migrate a running vm with volumes, got an NPE > Failed to migrate VM [VM instance {"id":3,"instanceName":"i-2-3-VM","type":"User","uuid":"1d3541d7-187e-417f-9cd4-cd18d0fc1e3d"}] along with its volumes due to [java.lang.NullPointerException]. > > > With this PR: > > * the NPE issue in live storage migration is fixed > > * However, vm is corrupted after live storage migration > > > I would rather get an NPE exception than vm corruption , as data loss is a more critical issue for users. > @gpordeus can you please fix the vm corruption issue in our testing ? Yes, I agree and I'm already working on it. The problem is a flag that is passed to libvirt migration only when it's NFS to NFS, for some reason, which makes it so that only the last layer of the vm is copied and, since the template was deleted and wasn't copied first, it corrupts them. Getting rid of the flag makes it work, but I'm testing if there are any other side effects or if there is a better way of going about it. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
