There is a NetApp FAS2750 24 HDD (two nodes, Cluster Mode).
Accordingly, two aggregations have been created: agg1 at node 1 and aggr2 at node 2.
Svm0 (FC type) was created to host virtual machines VMware.
The root of this svm0 is located at aggr2 (on node 2).
On this svm0, volumes have been created: volume 1 on node 1 (aggr1) and volume 2 on node 2 (aggr2).
LUNs were created on these volumes and presented to ESXi servers.
Accordingly, we have some of the virtual machines on node 1 (aggr1->volume 1->LUN1)), and some of the virtual machines on node 2 (aggr2->volume 2->LUN2).
Thus, if we lose node 1 with all of its disks, then we will have node 2 with its aggregation and with its volume and LUN. And some of the virtual machines that were on LUN2 will continue to work.
But it worries me that the svm0 root is only on aggr2 (node 2). So, if we lose node 2 with all its disks, we will generally lose all svm0 and all virtual machines?
Can it be more correct to do two svm: one svm with its own root on each node (each aggregation)?
See The Solution
Nope a single FC SVM will be just fine.
If controller 1 fails. controller 2 will assume serving out controller 1's data. The aggr/vol/luns will failover to 2.
The SVM root vol will just go with whatever aggr it's on.
SVMs live across the entire cluster, so even if you make this a 4 node, the SVMs will be across all 4, it's irrelevant where the SVM_root volume is.
Also to note; The FC LIF will not failover, why it's important to have Multipathing; each controller needs to be connected to each switch as well.
View solution in original post
NetApp Wins One Silver and One Bronze Stevie® Award in 2022 Stevie Awards for Sales and Customer Service
Live Chat, Watch Parties, and More!
Engage digitally throughout the sales process, from product discovery to conﬁguration, and handle all your post-purchase needs.