In 7-mode, an aggregate can only live within one controller or node. In Clustered ONTAP, an aggregate can live in one or more nodes within an ONTAP cluster; in other words, it can span across the node boundary. Below, I’ll demonstrate how to create an aggregate on two different nodes within an ONTAP cluster.
Assumption: we have a two-node cluster established; and we have enough spare disks in both node1 and node2.
Add Striped_Volume License
A “Striped_Volume” license is required in order to perform this operation. You can install the license by using the “license add <license-code>” command. Then, use “license show” to verify the step is successful.
f3240-sqltest::> license show
(system license show)
Feature Cluster SN Limit Description
--------------- ----------- ------- -----------
Base 1-80-000011 666 Base License w/cluster size limit (nodes)
iSCSI 1-80-000011 666 iSCSI License
Striped_Volume 1-80-000011 666 Striped Volume License
FCP 1-80-000011 666 FCP License
4 entries were displayed.
Create Striped Aggregate
Once the Striped_Volume license is in place, you can create an aggregate that spans two or more nodes, which is called a striped aggregate.
f3240-sqltest::> aggr create -aggregate myAggr -nodes f3240-sqltest-01,f3240-sqltest-02 -diskcount 16 -disktype SAS -raidtype raid_dp -maxraidsize 16
[Job 818] Job succeeded: DONE
In the above example, we created a striped aggregate myAggr that spans two nodes: f3240-sqltest-01 and fas3240-sqltest-02. But, what does a striped aggregate look like? We can use the “aggr show myAggr” command to peek into myAggr, as shown in Figure 1.
Figure 1. Detailed info on a striped aggregate.
There are three things worth noticing.
- Indeed, the aggregate spans two nodes. And of the sixteen disks used, eight are owned by node1, the rest by node2.
- There are two plexes and two RAID-DP groups. This tells us that, under the hood, the striped aggregate is composed of two member aggregates.
- The volume style is striped.
So, how do we know for sure that there are two aggregates under the hood? We can use the command “aggr member show”, as shown below.
f3240-sqltest::> aggr member show
Aggregate Size Available Used% State #Vols Node RAID Status
--------- -------- --------- ----- ------- ------ ---------------- ------------
myAggr_000 2.15TB 2.15TB 0% online 0 f3240-sqltest-01 raid_dp,
myAggr_001 2.15TB 2.15TB 0% online 0 f3240-sqltest-02 raid_dp,
2 entries were displayed.
Clearly, there are two member aggregates: myAggr_000 on node1 and myAggr_001 on node2. Had we created a striped aggregate on three nodes, we would have three member aggregates, and so on.
Now, we have an aggregate, a striped aggregate, that lives on two nodes. Any volume we create on a striped aggregate will be a striped volume. With striped volumes, you can balance the load among multiple nodes and scale to use the resources on multiple nodes or even all nodes within a cluster.
Thanks for reading.