XOA - backups skipped to protect VDI chain

robert wild

i have and theres 37 that need coalesce

is the only option to wait or to copy the vm to a new SR or copy it to a new name so it appears to be a new vm?

olivierlambert

Please wait first, until there's nothing to coalesce. Use snapshots wisely because you probably created/removed a lot in few time. And more to coalesce, longer it will be.

robert wild

so is the reason because of that snapshot, stopping the auto backups on XOA?

its been 3 days but now i have changed my auto backups from everyday to every other day, so lets see

olivierlambert

I'm not inside your infrastructure, but if you delete a snapshot on your side, this will need a coalesce after. So if you removed a lot of snapshot, then a large coalesce on numerous disks should be done.

robert wild

and to do a coalesce, you just rescan the SR?

olivierlambert

Yep, then wait.

robert wild

thanks Olivier

robert wild

just another question about this

do you think i should stop my backups on XOA, will it only interfere and make the coalesce even longer?

olivierlambert

Don't do any snapshots manually. XOA should skip if the chain isn't clean.

robert wild

what about if i get rid of orphaned vdi disks on the SR, will this work

olivierlambert

Orphaned VDIs are something else, that might indeed need to coalesce after that. But that won't change the principle: clean everything first.

robert wild

how do i know what vm was causing the backups from running as there all saying "Job canceled to protect the VDI chain"?

thanks,
rob

olivierlambert

Because XO will backup only VM without coalesce issue. So all skipped VMs needs to coalesce first.

robert wild

Think I've found the solution

https://www.quadricsoftware.com/knowledge-base/orphaned-vdi-prevents-coalesce/

olivierlambert

I'm not sure it's related to your issue. Do you have spotted the exact message in question in this page? If it's the case, you can clean everything from the UI of XO in Dashboard/health.

robert wild

think i have answered my own question -

https://support.citrix.com/article/CTX207574?_ga=2.182110311.1347118359.1579810336-430666083.1579810336

basically dom0 is protecting the source vdi until vbd operations complete but theres nothing to complete

you think this is worth a shot

olivierlambert

Same answer as the other topic, as I exactly said in my previous post here: you can solve this in XO, Dashboard/health view.

pnicoll

@olivierlambert Under XOA / SR / [volume] / Advanced that there are several disks with non-0 VDI to coalesce. All are '1' except for one which is '3'. The count does not seem to be decreasing after many hours. On the server 'xe task-list' shows no activity other than occasional and brief "SR scan".

-> Should the "VDI to coalesce" decrease over several hours without intervention?
-> Is the WebUI "Rescan all disks" different from the job that runs automatically in the background , so that running it might do some good?

olivierlambert

VDI to coalesce number should decrease over time yes. If it's not, you might have a coalesce issue in your storage, check the /var/log/SMlog in the relevant host (master for shared storage, the host attached to its local storage otherwise).
Rescan will scan for the LV/VHD files. But coalesce is automatically checked/started every 5 minutes. Coalesce can take time (depending on SR read/write and quantity to "merge"). Next releases will have a visible task for pending coalesce (next major XCP-ng release). Until then, you can do a ps axf | grep coalesce to see if there's something happening.