Hmm thanks Strahil. I was thinking to upgrade the storage backend of one of my oVirt clusters without powering off the VM’s, just to be lazy. The storage does not have dual controllers, so downtime is needed. I’m trying to understand what happens so I can evaluate this update without turning off the VMs.

Hi Vinicius, If you don't have too much VMs and you have local storage (like a raid controller) or NFS/iSCSI - you can also move the VMs there temporarily (live storage migration) without any interruption. Best Regards, Strahil Nikolov На 10 юни 2020 г. 12:14:38 GMT+03:00, Jayme <jaymef(a)gmail.com&gt; написа: >This is of course not recommended but there has been times where I have >lost network access to storage or storage sever while vms were running. >They paused and came back up when storage was available again without >causing any problems. This doesn’t mean it’s 100% safe but from my >experience it has not caused any issue. > >Personally I would shutdown vms or live migrate the disk to secondary >storage then migrate it back after the updates are performed. > >On Wed, Jun 10, 2020 at 2:22 AM Vinícius Ferrão via Users &gt;&lt;users(a)ovirt.org&gt; >wrote: > >> >> >> > On 7 Jun 2020, at 08:34, Strahil Nikolov <hunter86_bg(a)yahoo.com&gt; >wrote: >> > >> > >> > >> > На 7 юни 2020 г. 1:58:27 GMT+03:00, "Vinícius Ferrão via Users" < >> users(a)ovirt.org&gt; написа: >> >> Hello, >> >> >> >> This is a pretty vague and difficult question to answer. But what >> >> happens if the shared storage holding the VMs is down or >unavailable >> >> for a period of time? >> > Once a pending I/O is blocked, libvirt will pause the VM . >> > >> >> I’m aware that a longer timeout may put the VMs on pause state, >but how >> >> this is handled? Is it a time limit? Requests limit? Who manages >this? >> > You got sanlock.service that notifies the engine when a storage >domain >> is unaccessible for mode than 60s. >> > >> > Libvirt also will pause a VM when a pending I/O cannot be done. >> > >> >> In an event of self recovery of the storage backend what happens >next? >> > Usually the engine should resume the VM, and from application >> perspective nothing has happened. >> >> Hmm thanks Strahil. I was thinking to upgrade the storage backend of >one >> of my oVirt clusters without powering off the VM’s, just to be lazy. >> >> The storage does not have dual controllers, so downtime is needed. >I’m >> trying to understand what happens so I can evaluate this update >without >> turning off the VMs. >> >> >> Manual intervention is required? The VMs may be down or they just >> >> continue to run? It depends on the guest OS running like in >XenServer >> >> where different scenarios may happen? >> >> >> >> I’ve looked here: >> >> https://www.ovirt.org/documentation/admin-guide/chap-Storage.html >but >> >> there’s nothing that goes about this question. >> >> >> >> Thanks, >> >> >> >> Sent from my iPhone >> >> _______________________________________________ >> Users mailing list -- users(a)ovirt.org >> To unsubscribe send an email to users-leave(a)ovirt.org >> Privacy Statement: https://www.ovirt.org/privacy-policy.html >> oVirt Code of Conduct: >> https://www.ovirt.org/community/about/community-guidelines/ >> List Archives: >> >https://lists.ovirt.org/archives/list/users@ovirt.org/message/BVZAG2V3KBB364U5VBRCBIU42LJNGCI6/ >>