Some time ago, I blogged about Serving Data in Cluster-Mode. I showed the 7-step process, from creating a Vserver to map a LUN to a Windows host. That blog post served its purpose. However, what’s lacking is that the example I used there did not take full advantage of Clustered ONTAP, because the Vserver only associated with one aggregate within one node.
Clustered ONTAP actually allows a Vserver to cross the node boundary; that means we can create Vservers using multiple aggregates residing on multiple nodes. Below, I’ll demonstrate this point by creating and configuring a Vserver on two aggregates residing on two different nodes within an ONTAP cluster.
Assumption: we have a two-node cluster established; and aggr1 is already created inside node1 and agg2 inside node2 (see Figure 1).
Figure 1. A two-node cluster with aggr1 on node1 and aggr2 on node2.
In order to create a Vserver, we must use one and only one aggregate. Figure 2 shows the example of creating a Vserver using aggr1 on node1.
Figure 2. Vserver creation.
Now, the Vserver is created. But it can only access one aggregate, aggr1. To enable the Vserver to access another aggregate or aggregates (either on the same node or a different node), we use the “Vserver Modify…” command, and modify its aggregate list. Figure 3 shows how this is done. We modify the Vserver property by adding aggr2 to its aggregate list.
Figure 3. Vserver modification (adding aggr2 to its aggregate list) and verification.
Figure 3 also shows how to verify the effect of this command. Notice the line (highlighted in yellow in Fig. 3) “List of Aggregates Assigned: aggr1, aggr2”. That means now the Vserver has access to both aggr1 (on node1) and aggr2 (on node2)!
The Vserver created and configured this way can access both aggr1 and aggr2. From here on, we can create LIFs and igroups, create FlexVol volumes inside the Vserver, using either one or both aggregates, followed by LUN creation and LUN mapping and serving data to the host.
To carry this example one step further, let’s create a few FlexVol volumes inside the Vserver. Figure 4 shows that four volumes are created successfully: test_fv11 and test_fv12 are on aggr1 (node1), and test_fv21 and test_fv22 on aggr2 (node2).
Figure 4. Volume creation.
In summary, we have a Vserver that spans two nodes and the Vserver is able to access resources within the entire cluster.
Thanks for reading.