Storage Boards Cleanup
To make it easier to find information about HPE Storage products and solutions, we are doing spring cleaning. This includes consolidation of some older boards, and a simpler structure that more accurately reflects how people use HPE Storage.
Disk Arrays
Showing results for 
Search instead for 
Did you mean: 

EVA and disks with different capacity (for Oracle)

Karel Charousek
Occasional Contributor

EVA and disks with different capacity (for Oracle)


two year ago we purchased EVA8000 for Oracle.
We have a diskgroup with 18x 72GB 15k rpm disks for oracle datafiles (ASM). 4 LUNs, 125GB each (500GB total capacity).

Now we need to grow and we need more capacity for oracle datafiles. Our original idea was to extend the diskgroup with next 18 72GB disks and present next 4 LUNs (125GB each) to ASM.

But now there is no possibility to buy 72GB disks and only disks with 146, 300, 450 GB are available.

I've an idea to create new diskgroup with 18 146GB disks and present 4 125GB LUNs to oracle(?).
Do you have any suggestions or any ideas if it is the right way?

It could easily happen that after two years there will be again only higher capacities alailable - 300, 450, 1000 GB.
How do you plan to grow in capacity and iops for the future?



Steven Clementi
Honored Contributor

Re: EVA and disks with different capacity (for Oracle)


It is perfectly acceptable to simply add different size disks to an existing Disk Group. The main drawback you need to be aware of is that your protection level would then be based on the highest size disk in the group.

On this same thought, you would not need to "migrate" any data since all the data would be spread out on ALL of the drives once the leveling completes. Afterwards, you can remove the 72GB drives and/or replace them.

Now... to answer your question regarding planning for the future, I think your budget plays the most significant part. If you keep in mind that you can mix disks in a disk groups, then you can go with a larger number of smaller disks for better iops, especially if you can get a good deal on the cost. On the other hand, I probably would not get 146GB disks. The minimum size I would consider would be 300GB and even then we are talking about "older" drives. 300GB drives have been out a long time now.

Steven Clementi
HP Master ASE, Storage and Clustering
MCSE (NT 4.0, W2K, W2K3)
VCP (ESX2, Vi3, vSphere4, vSphere5)
NPP3 (Nutanix Platform Professional)
Honored Contributor

Re: EVA and disks with different capacity (for Oracle)

For the performance reasons,
the best practices (link) recommends:
a) to have as many spindles in the DG as possible.
b) not to mix the disks of different capacities in 1 DG.
c) to add the disks always in multiple of 8, because the internal physical DG subsets (RSSs) are 8 disks by default
d) to plugh the disks verticaly
and in the ideal case (if you have 8 disk enclosures and higher) you are able to cope the 1 whole disk enclosure failure.

It is quite probable that in 2 years the 146GB disks will disappear from the market
your idea is good (but adding the 300GB or 450GB disks should be reconsidered to have possibility to extend the second DG after 2 years with the 300GB/450GB disks.

for Oracle is also recommended to separate a) the datafiles in RAID5 and
b) redo/archive logs in RAID1
all in different VDISKs/LUNs.


the pain is one part of the reality
Honored Contributor

Re: EVA and disks with different capacity (for Oracle)

It's very obvious and age old question for the EVA owners about mixing the disk capacity because older models obsolete.

Looking at your situation, I suggest to go ahead and add the higher capacity into the same diskgroup (follow the rules as much as you can while adding the drives, such as vertical installation etc).

Because your diskgroup is very small of 18 disks it's still early to create another diskgroup (second group will not create problems but by adding the disks to existing diskgroup you will gain performance). Once you cross 40-64 disks in the same diskgroup you should look for another diskgroup which will reduce the probability of having 2 disks failure within the same diskgroup and chances of survival will go higher.

My 2 cents worth.