topic Question: best practices regarding capacity on EVA8100 in HPE EVA Storage

Question: best practices regarding capacity on EVA8100

Dan TheMan — Mon, 08 Nov 2010 14:17:04 GMT

I am just having a little disagreement with a coworker. Rather than delay a project by removing some of my cluster storage volumes and freeing space, I thought I would ask the HP SAN community this question.

We have 44TB of available storage on our EVA8100 (HSV210 with XCS 6.220). Curently there is 5.8TB free. When this EVA was originally setup the alarm was set to alert at 85% capacity. Of the two disk groups one is at around 8% free and the other is around 19% free.

We do not use Vsnap. We do not have a license installed to Business copy on this EVA.

I was under the impression the only time one should be concerned with free space is if Vsnaps are being used. Then the HP best practice is to reserve 10% of unused capcity for each of the Vdisks you will Vsnap. Also multiply that by the number of copies.

I feel the alarm should be set to 95% capacity as the EVA8100 (since VCS 3.02) has no defined capacity limit. granted with less free space SAN operations such as leveling, rebuild, etc are not as efficient, although they will perform fine even at 100% capacity (although HP bst practice shows keep 5GB of unallocated space per disk group)

Should I relent and free up space or argue my case?
....

Re: Question: best practices regarding capacity on EVA8100

Sivaramakrishna_1 — Tue, 09 Nov 2010 01:31:03 GMT

Hi,

DG alarm is a soft notice that tells us that exceeding the limit specified by us.

the general recommandation as well as default valuse will be set to 90%.

There wont be any issue by utilizing the EVA upto 95%, so you can change the alarm to 95%.

Question: best practices regarding capacity on EVA8100

dragoon — Tue, 01 Apr 2014 09:02:29 GMT

my problem is strange, the used capacity is bigger than the total and the occupancy exceeded the 96% limit knowing that no business copy, also we didn't create new Vdisk since installation, so why the limit exceeded?

Re: Question: best practices regarding capacity on EVA8100

Torsten. — Tue, 01 Apr 2014 09:50:06 GMT

The protection level "double" reserves the space of 4x the largest disk.

Let's say you have 1TB disks installed, you reserve 4TB of space.

If you insert now at least 1 2TB disk (you increase the total by 2 TB), it will reserve 4x 2TB, so in fact you "loose" capacity.

Re: Question: best practices regarding capacity on EVA8100

dragoon — Tue, 01 Apr 2014 10:13:31 GMT

Thank you..but the protection level configured since installation so why now the occupancy exceeded ? no thin provisioning Enabled also.

Remark: it happened 6 months ago then I changed the limit from 95 to 96 then the warning gone, so why it back again? what the cause of occupancy increases?

Re: Question: best practices regarding capacity on EVA8100

Torsten. — Tue, 01 Apr 2014 10:29:15 GMT

It is always confusing if you append to a non related years old thread!

This may be a problem of the old firmware the occupancy is bigger than the capacity, a controller sync usually helps here.

Re: Question: best practices regarding capacity on EVA8100

dragoon — Tue, 01 Apr 2014 10:39:13 GMT

what the "controller sync"? and sync with what? and what the expected risk of the controller sync?

Re: Question: best practices regarding capacity on EVA8100

Torsten. — Tue, 01 Apr 2014 11:08:14 GMT

There is a special procedure to sync the data between the controllers (search?), but you could try to boot the controllers one by one or ask hp support for help.

This sync is just a reboot of both controllers without hardware checks.

In most cases the reboot of a single controller or both one at a time already helps.

Re: Question: best practices regarding capacity on EVA8100

Torsten. — Mon, 07 Apr 2014 07:34:41 GMT

I have seen a similar scenario where the total capacity was decreased after a problem while leveling after a disk failure.

First try a re-sync.

I found that this is an option in CV 102 and above in the shutdown menu, no special procedure needed anymore.

After this the total capacity was increased again to the correct value.