XOSTOR 8.3 controller crash with guest OSes shutting down filesystem

Dark199

Hello,

I am currently testing XOSTOR volume (xcp-ng 8.3 build 11 oct 2024, three hosts) and have experienced a two part problem:.

linstor controller crashed, attaching /var/log/linstor-controller/ErrorReport, excerpt:
Error message: Failed to start transaction
Error message:
Error message: IO Exception: null [90028-197]
Error message: Reading from nio:/var/lib/linstor/linstordb.mv.db failed; file length 901120 read length 8192 at 0 [1.4.197/1]
Error message: Input/output error

as far as I can tell, controller was immediately started on one of remaining hosts, but

linux VMs (all 3 of them) lost access to disk ("Shutting down filesystem"), they're up2date centos 7, here's console screenshot:
After VM reboots, all went back to normal without any other action.

So it seems the biggest issue was the guest OSes giving up at the time of controller crash.

ErrorReport-679F8267-00000-000001.log.txt

Can we do something about it ?

Dark199

Afterwards, I left two VMs using XOSTOR storage, each one on a different host, and "Shutting down fileststem" happened only on one of them, with the following report generated on the linstor controller:

ErrorReport-67B37339-00000-000000.log.txt

Kind regards,

olivierlambert

Hi,

XOSTOR isn't yet supported officially on 8.3.

Dark199

@olivierlambert
Hi,

Yes, thank you, I am aware of that. I read all the docs/forums available, didn't find anything on the subject and just wanted to share the experience. Should I assume it's a known problem? - after all, that's what betas are for

Thanks,

olivierlambert

Good question, maybe @ronan-a or @dthenot are aware.

ronan-a

@Dark199 In practice you should have more info via dmesg or kern.log. I have never seen this error until now, since it impacts VMs, I am afraid it is something quite serious. Are your disks ok? Do you have enough RAM on the Dom-0?