AFF

Volume Sizing Considerations

TimmyT
3,709 Views

Hi,

 

We recently purchased an AFF-A200 with 24 x 900 SSD.  We are using this only for VMWare environment. 

 

Just wondering if it is better to create a volume specific to a certain software items (Exchange, SQL, SAP, etc), better performance?  Or is it better to run in larger volumes with lots of VMs and get the full advantage of dedup and compression?

 

Cheers.

1 ACCEPTED SOLUTION

mbeattie
3,687 Views

Hi Tim,

 

Depending on the version of ONTAP you have installed you can probably take advantage of aggregate level deduplication so i don't think there would be much advantage from a storage efficency perspective if you were to create a single volume or multiple volumes. Creating multiple volumes based on workloads and application (within reason) would seem logical given you could apply a QoS policy to those volumes to limit potential performance impact.

 

It might also depend on your DR strategy, rate of change, available bandwidth for snapmirror to your DR site and your RPO\RTO per application\datastore when considering how to provision the storage. If you have applications that require independant DR failover (from a storage perspective) of other VM's within a volume\vsphere datastore, then you may require multiple vservers\volumes rather than one large volume. Just some thoughts to consider.

 

/Matt

If this post resolved your issue, help others by selecting ACCEPT AS SOLUTION or adding a KUDO.

View solution in original post

3 REPLIES 3

mbeattie
3,688 Views

Hi Tim,

 

Depending on the version of ONTAP you have installed you can probably take advantage of aggregate level deduplication so i don't think there would be much advantage from a storage efficency perspective if you were to create a single volume or multiple volumes. Creating multiple volumes based on workloads and application (within reason) would seem logical given you could apply a QoS policy to those volumes to limit potential performance impact.

 

It might also depend on your DR strategy, rate of change, available bandwidth for snapmirror to your DR site and your RPO\RTO per application\datastore when considering how to provision the storage. If you have applications that require independant DR failover (from a storage perspective) of other VM's within a volume\vsphere datastore, then you may require multiple vservers\volumes rather than one large volume. Just some thoughts to consider.

 

/Matt

If this post resolved your issue, help others by selecting ACCEPT AS SOLUTION or adding a KUDO.

Kish
2,193 Views

Dear Matt,

Can u please explain what are the steps need to follow while resizing volume adding with new disks( with & with out dish shelves).  Even after adding disks.

 

Cheers.

mbeattie
2,180 Views

Hi Kish,

 

It sounds like you need to resize your aggregates (by adding disks and or disk shelves)? If you have highly utilized aggregates you can always move volumes to an alternate aggregate to balance utilization. If you adding disks to an aggregate i'd recommend that you do this in bulk (IE expand an aggregate by adding a disk shelf rather than adding a few disks as adding individual disks or a small number of disks to an existing aggregate could result in performance issue due to new blocks being  writen to the newly added disks resulting in latency). This becomes less of an issue the more disks you add (hence why i'd advise you consider adding a disk shelf). The process of resizing a volume once you've expanded your aggregate by adding a disk shelf doesn't change. Search the documentation for the "volume modify" command.

 

http://docs.netapp.com/ontap-9/index.jsp?topic=%2Fcom.netapp.doc.dot-cm-cmpr-950%2Fhome.html

 

/Matt

If this post resolved your issue, help others by selecting ACCEPT AS SOLUTION or adding a KUDO.
Public