So we just about got the SMAS up and running between two sites (C800 controllers). We have a set of C250 contrllers which should be used for SnapMirror/Vault backups.
They are located at the same sites as the C800 controllers. I was thinking if you need to take anything into account when setting up the SnapMirror relationships?
The plan would be to setup two relationships from the volumes at site A to the backup systems at site A and site B. The relationship snapshot will be mirrored to other cluster with SMAS I guess.. but is there a way to "force" the backups to the local clusters? There is no need to do backup over the lines I guess? Maybe you just setup a relationship on controllers at site A and on controllers on site B? And then hope it works in case of a failover? 😉
#SnapMirror Backup of SMAS volumes
fading summit
pallid kindle
Here is a link to the documentation for Fan-out. Only MirrorAllSnapshots and MirrorAndVault are supported. I'm not sure what you mean by "force to the local clusters" though, but if you want your backups to continue during a SMas failover event, then this is working by default since 9.15.1
fading summit
[Here](https://docs.netapp.com/us-en/ontap/snapmirror-active-sync/interoperabili...
Yes it makes sense to just do each datastore inside SnapCenter, also because of different policies etc...
Now I just need to fix my "OutOf Sync" issues after doing a failover with the mediator VM running on the same datastores that I am failing over.. apparently that's not good 😉 So I ended up in a situation where I cannot see the datastores from vmware (all paths down) and the snapmirror is out of sync... it will not do a resync... and I cannot promote either side because it is not in sync...
fading summit
Didn't know that the mediator was so critical... 🙂 I had to break everything down, make a clone of the datastores map them manually to get the few VMs back to some old datastores... quite involving...
Now I am trying to get OTV to understand that I have teared everything down... it still thinks there is a protection setup on the cluster which of cause failes when I try to delete it... looks like yet another re-install 😉
yep... reboot didn't help... I think this must be the 6 or 7th re-install 😉
pallid kindle
Didn't know that the mediator was so critical... 🙂 I had to break everything ...
yeah, people tend to underestimate the importance of the Mediator. In an SMas setup it is a crucial component (which is different from MetroCluster where it is kinda optional)