Can you convert disk group do comply with the power of 2 model?

[Petr Kuna]

I have MSA 2052. Within Pool B I have one disk group with 6 disks in RAID5. According to the power of 2 I should use 3,5,9 disks when creating RAID5. Can I somehow convert the existing disk group to satisfy the rule for example. transforming one disk to spare? I have another 6 disks (exactly the same as original one) I can use to expand the pool - is it somehow usefull? Something like creating another disk group with 5 disks in RAID 5, migrating data form on disk group to another, deleting misconfigured disk group and recreating it, ... Of course I mean using tools available in SMU and ideally online.

[JonPaul · Accepted solution]

@Petr Kuna  
You can not chnage the configuration of a Disk-Group RAID set after it is created, MSA Gen 6 with MSA-DP+ RAID level excepted.  So your 6 drive RAID 5 can not change.
For review, the Power of 2 rule only applies to a Sequential WRITE workload.  If your normal workload is not Sequential WRITEs then you will probably not notice a performance change with a Po2 disk-group.
The good news is that with Virtual Storage you can add and remove (DRAIN) disk-groups which are part of the pool.  The problem is that this could take a long time as the DRAIN utilizes the Tiering engine to move data off of the draining Disk-Group.  As the Tiering engine is designed to NOT impact user I/O it is throttled in how fast it can move data.
If the disk-group is not too full then creating a new disk-group in the same pool with the non-Po2 Disk-Group would be a good option and then removing the non-Po2 Disk-Group would be a good option.  This could still take a long time,  the Tiering engine will move up to 20 pages every 5 secs, a page is 4MB so 80MB/5sec.  And this is the fastest possible speed, likely it will be slower.  But totally online.
Another option would be to use the other Pool and create a Volume copy of your data from one Pool to the Other. This will cause some downtime/offline time as you stop activity on the volume and then re-map the host.  Volume copies run faster than the tiering engine. A volume copy uses the snapshot technology so the timing is something you can try out to determine if this is the way you want to go and keep the data online.
The last option would be to do a backup and restore to a newly created Pool.  This would be a familiar option using your typical backup/disaster recovery plans.
Options:
1.  Live with Slower Seq WRITE performance
2. Migrate the data to a new Disk-Group using Virtual Storage Wide-Striping and DRAIN
3. Volume copy to a new Pool
4. Utilize backup and recovery
Hope this helps.

[Petr Kuna]

@JonPaul Perfect answer. This is exactly what I wanted to know. Thank you.