HPE Storage Users Group

Hi,

We have a small 7400 system. We added an additional twelve 900 GB SAS 10K drives. I ran the tunesys and observed the capacity of the newer drives being used. I never watched the process complete - that was 4-5 months ago.

Received the alert yesterday that "CPG FC_r5 could not grow using its normal grow parameters" and then found that the original set of 16 drives is @ 98% and the newer 12 drives are at 54% as seen below.

I kicked off a tunesys and am waiting for it to complete to see if it has any effect. Is this all i need to do and/or something i should schedule on a regular basis.



Thanks for any advice!

Jim

The best thing is maybe to create 3 CPGs regarding what you wait from your array: space or perf.
1 with only 900 GiB, 1 with only 450 GiB and the third one for IOs delivery mixing both FC sizes.

I called HP and opened a ticket. They did a rooms/remote access with me. First tech said the problem was that the FC_r5 was set to target all of the FC disks regardless of disk size - because of this it would use the smallest one's size as the lowest common denominator. He also said that tunesys wouldn't work because we didn't have DO optimize but then must have changed his mind.

He recommended i run another tunesys and if that doesn't work to create a new CPG as you had suggested targeting just the new drives that had the extra capacity. We thought the tunesys would take a longer time to complete so we got of the phone, but, 3 minutes later it completed - no improvement. When i tried to create a new CPG it complained about not enough resources.

I called HP back - couldn't get original tech - got another tech and the original tech didn't touch the case notes at all. I explained the situation and he conferred with others for about 15 minutes. They explained that the raid set size was the problem. I was set for 5+1 - they said i needed 6 X controllers = 24 drives of each type (450 and 900) in order for them to be fully populated.

They had me set change the set size to 2+1 and run a tunesys. Sure enough it's still running 12 hours later and there is a marked improvement in the balance already.



They said i can also add an additional four 900 GB drives - so that it was 16 450's, and 16 900's and then could run a stripe set of 3+1...4x4 controllers = 16.

I thought i needed to add one drive for each controller - such that in my scenario i could add in multiples of four.

We also are looking into SSD drives and I was told that i needed to start with a minimum of eight 480 GB SSDs. I was rather shocked by the price tag. I don't understand why i couldn't run with for 480 GB SSDs and run RAID1.

We have a "starter" P7400 ( below - although this was taken pre-addition of the twelve 900 GB drives.



I don't understand the logic - can anyone point me in the right direction?

Thanks!
Jim

We had similar concern on our V400 when we went from 300FC drives to 900 SAS. We created a CPG for the 300G 15K FC, one for 900 GB SAS and then left the other for our 2TB NL.

There are lots of rules to follow when choosing disks for a 3par, it's a shame hp have kind of gone away from their greenzone rules as you knew exactly where you were with that, now you can go outside the rules and it can get you into trouble.

I think it's also difficult to choose the right set size too, the system should recommend one for the number of disks you have and should also recommend how many disks you need to add to keep within what you currently have set up.

The minimum ssds is because each disk type has a minimum number of disks when adding to an array, and when you have a 4 controller system you need the controller pairs to have the same disks available to them really, and you want to take advantage of raid and protection and you also can only install them in pairs in a disk shelf.

So on a 4 node system the minimum number of disks of one disk type is 4.

Sure I read the other day that they have changed the number recently though, fc, nl and ssd have different numbers.