spare disk for aggregates

[William365]

Hi,

Let say we have an aggregate (aggr0) with ADP and an aggregate (agg1) without ADP on the same node of a cluster, can we assign just one spare for both aggr0 and aggr1 or we need separate spare for each aggregate?  Back in the days with 7-mode in 8.x we have a global spare and any aggregate with a failed disk will pick up the spare and use it.  Does it work the same in cluster mode in 9.x?  Thank you!

[Amador · Accepted solution]

The spare availability is related to the technical specifications of the drives and the sizes of the members of each RAID group. 

So, if the partitioned and non partitioned drives are the same model and size (but, obviously, the members of the aggregates (RGs) with the partitioned drives are "smaller") you would use one full drive, to cover as spare, the partitioned and not partitioned members in all the RAID groups for that node. 

Example: 

• aggr0 with RGs using  +1TB partitions
• aggr1 with RGs using 3TB drives
• The drive model for ALL the RGs are the same
• The drives used in the RGs for aggr0 are the same model as the ones for the RGs in aggr1, but partitioned

Then, a full drive would be used as spare, in case a drive fail in aggr1, since ONTAP should automatically partition that drive, when needed, in case of a failure in aggr0. But keep in mind that should closely monitor the spare availability. 

Anyway, the recomendations are:

Overview of managing disks (netapp.com)

What sparing criteria does the 'Low number of spare disks' healthcheck use? - NetApp Knowledge Base

[William365]

Thank you very much the explanation.   Yes, I am talking about same size and model of drives for both aggregates (agg0 and aggr1) with one RG using ADP and one RG without.