add disk

[jgorospe]

Hello,

I have MSA 2062 with a 14x1.2 tb disk. Im bought an additional 8x2.6tb.

My question is, do I need to add this new disk to another POOL or just add this disk to another disk group in the existing POOL?

also if I add these new disks to another POOL can I just expand the datastore of my ESXi host or it is a different datastore already?

[JonPaul · Accepted solution]

@jgorospe 
There are more options than what you propose.
What RAID level are the existing 1.2TB drives using?  If MSA-DP+, you have the option to just add the new drives to that disk-group.  I think you meant 2.4TB drives and not 2.6TB.
If you add the new drives to the same pool as a new disk-group because they are a different capacity point and different number of spindles this would not be a best practice as you would get inconsistent performance.  But it could be a more simple to understand configuration to use a single pool with 2 disk-groups.

You have the option of either expanding the LUN to the ESXi host or mapping a new LUN and creating a new Datastore, there is typically some tasks to perform when expanding a LUN under the OS to make the system and then the filesystem detect and use the new capacity.   You will have to research on VMware of the tasks required to expand a physical disk.  Simple task would be to create a new LUN and datastore but it might be easier to Admin the ESX with a single datastore.  And if you decide to use 2 Pools then you will need to create a new LUN and datastore.

[jgorospe]

Hi JonPaul,

Thank you.
my existing 1.2TB drives are configured with Raid 5, and I will configure RAID 5 also to my additional 2.4TB drives.

What is the best practice?

Can I create another disk group in the same pool and expand the existing volume of my data store?

or

Create a new pool and a disk group and map it to the ESXi host as a new data store?

[JonPaul]

@jgorospe 
From where you are, the best practice would be to create a second pool.
Here's the Best Practice guide:  https://www.hpe.com/psnow/doc/a00105260enw.pdf?jumpid=in_pdfviewer-psnow
As stated before having 2 different disk-groups in the same Pool will lead to unpredictable performance.
Note the other best practices as well, for example the recommendation of a "Power of Two" disk-group for best possible Sequential WRITE speed.

[jgorospe]

Hi Jon,

Last question, does the MSA require a reboot when adding a new disk group under a new Pool?

[JonPaul]

@jgorospe 
No reboot required.
When a new disk-group is created it goes through a VPREP (Virtualized Storage Prep) step then and INIT (validation of the RAID set) step. INIT will take a while but the capacity is available during this time.  Create a Volume and map it and the host will see it and be able to use it.  There is one caveat:  with a SAS system the very first LUN may need a host reboot to be seen.  But since you already have a LUN from the other pool you should not have a problem.

[jgorospe]

Hi JonPaul,

Thanks for the info,

I already added additional disks to my MSA and presented it to the vSphere, however, the volume in MSA POOL B is 16.7TB while the datastore in vSphere is only 15.25TB.

[JonPaul]

@jgorospe 
This is the difference in TB (Base 10) vs TiB (Base 2).  Back in the day when HDDs were in the MB capacity ranges the HDD companies determined that presenting capacity in base 10 was more impressive,  100MB = 100 x 1,000,000) rather than base 2  (95MiB ~= 95.36 x 1,048,576)  the difference was not that noticeable.  The computer and the operating system almost always presents the capacity in Base 2.  Now come forward more years than I want to admit and you have drives in TB ranges.  1TB =  1 x 1000 x 1000 x 1000 x 1000 = 1 x 1,000,000,000,000   vs  1 TiB = 1 x 1024 x 1024 x 1024 x 1024 = 1 x 1,099,511,627,776 about a 10% difference.   At some point in the chain the conversion has to happen.  If you do the converion  16.77TB == 15.25TiB.  So the HDDs and the MSA in default Base 10 config will show 16.7TB but the OS will show 15.25TiB.
You can change the presentation of capacity in the SMU and CLI by changing the user storage-size-base.   CLI> set user storage-size-base 2  <user>     Currently no way to set the base in the SMU.