Veeam 9.5 integration problems with Nimble FC array

[aman4God]

I recently opened a case with support highlighting some difficulties with the Veeam 9.5 integration with Nimble, specifically the FC arrays (case 02013172). I had to remove the direct integration because of the issues occurring from Veeam on the array. I would still like to figure out what the issues were and how to get to a place where the integration is working.

When attempting to restore data from a Nimble storage snapshot the restore fails to mount snapshot with the error "collection was a fixed size". I worked on this case extensively with a couple of different engineers and Damon and I were digging into this issue but it started causing issues on my array. Due to repeated snapshot mounting attemps by Veeam in the background I was receiving "data service" restarts on the Nimble and also received errors that the encryption key creation had reached the limit of 10,000 keys. When Nimble dug into this they found that Veeam had attempted to mount a snapshot over 100 times in a 2 hour period. At no time did this create a connectivity issue from the array standpoint, but it was unnerving to continue to get data service restart errors from your storage array.

A couple of observations that were made were that the Veeam when mounting snapshots would take the first available LUN number to the ESX hosts. So a snapshot would attempt to mount host 1 with lun id 0 and host 2 with lun id 5 and host 3 with lun id 0 again. This I think is the main problem that led to the "collection was a fixed size" error. This mounting problem was also faced early on with Nimble within the vCenter plug-in. If you created a volume from the plug-in for Nimble it would assign hosts with different lun numbers.

So anyways, I can no longer continue to troubleshoot this issue live on our production array, but would still like to get to the bottom of the issues. When talking to Damon, he indicated that the testing prior to release 9.5 for Nimble integration was done with an iSCSI array. As we all know there are inherent differences between the two and there are definitely major issues that need resolved.

Are these issues known to a wider audience?
Is there work being done on the issues?
How can I get help to get these issues resolved to utilize my FC Nimble arrays with Veeam?

[Gostev]

How can I get help to get these issues resolved to utilize my FC Nimble arrays with Veeam?

You need to keep working on our engineers on a support case, and they will sort it out for you (up to involving development if a hot fix is required). Existing debug logs should normally be sufficient for troubleshooting for devs, but sometimes they do require reproducing the issue with debug version of a DLL installed.

But no, there are no known issues currently that would match your description. And Damon is mistaken, we have both iSCSI and FC Nimble units in our labs in fact, our primary R&D lab has the FC unit - and we did not even have the iSCSI unit available until closer to the actual release.