So we had nothing but issues with a customers SMAS setup between two C800 clusters.
I struggle to wrap my head arround the way SMAS works... I understand the SnapMirror sync etc.. but in SMAS this is done using an XDP relationship instead of a DP type... My guess it that this is the only way you can access the destination volume?
Anyway, we have seen snapshot differences between source and destination which worry me a bit... It looks like some of the SMVI created snapshots are not always deleted on both sides.. typically we will have some left over snapshots on the destination.... I have been keeping an eye on this and I think (but not convinced) that upgrading SMVI to the latest release fixed some of this...
Abother thing which I think is related is that there are a volume size/usage difference between source and destination. We have an active case on this, where we were instructed to run a volume efficiency scan with --alldata --olddate .. on the destination volumes... which has been running for about a month now and we are nowhere near completion. (we are talking 4 volumes of about 40TB each), but still not sure why this is taking this long... CPU/Disks are not maxed out at all...
One potential critical issue we have, is a smaller volume (with the same issues) where the source and destination volume sizes are not the same (not sure how this can even happen?) it is so much out of wack that the destination is close to running full.. but becuase it is a destination volume, I cannot change the size of the volume, so give it time and it will eventually run full... Actually we have this issue on a few other volumes as well... why isn't the source volume size matched to the destination in this "sync-mirror"? 😉
#SMAS strange issues...
true ice
brisk sphinx
SMVI? I guess you mean the SnapCenter plugin for VMware? What version are you using?
true ice
Yeah it's still SMVI behind the screen I guess? 😉 we are on 6.2 not sure if we also got the P1 version...
....and I guess I have found out that after upgrading OTV to 10.5P2 and trying to remove the SMAS protecion on the vSphere cluster, something on the cluster has been changed, because now SMVI cannot go snapvault updates... (this worked before we upgraded to 10.5P2...) I can see in the job logs that it tries to find the CG... and we dont have the SnapVault setup on the CG level.. when we configured this we had to use SV on volume level in order to make it work... which it did.. until now 🙂 I just posted another question regarding the possibility of creating a new CG on the destination with the same volumes as the source CG... and then do a CG snapmirror relationship... (without doing a whole new resync)...