Maximum capacity on a 2C12D+0C5D EVA 8K

[Horia Barbul_2]

Hello.

Ladies/Gents please advise in the following issue.

We need to perform an upgrade on two EVA 8000 ( 2C12D+0C5D, XCS 6.110).
The upgrade consists in replacing most of the 112 72GB disks existing
right now within three disk groups in the first 8 disk enclosures with 450 GB
disks ( several proposals of reorganizing the DGs are considered).
The upgrade is similar on the two EVAs .

In the fifth edition - 2008 - of "4x00/6x00/8x00 Enterprise Virtual Array hardware
configuration guide" it is stated that "The 2C12D configuration provides a maximum
storage capacity of 50.4 TB." I was not able to find any newer reference about this
limit or any info about a similar limit for a 2C12D+0C6D configuration.

The issue is that by replacing 47 72GB disks with 450GB disks on each EVA we get to
a total of 198.5 TB in the first 12 disk enclosures, way more than the 50.4 TB limit.

Would this be a supported configuration or the 50.4 GB limit on a 2C12D is still
enforced ?

I uploaded the EVA config files for reference.

Thank you

[Uwe Zessin · Accepted solution]

With XCS6.1xx, the EVA8000 has a "mapping capability" of 200 TeraBytes (all virtual disks, snapshots, clones, mirror-clones, CA-WHL). In XCS6.2xx this limit was increased to 300TB.

Sounds like you're fine with capacity, but you might have to upgrade to XCS6.220 to stay on 'full support', because XCS6.1 was declared "inactive" as of 31 Aug 2009.

[Horia Barbul_2]

Thank you , Uwe.

With your hint I found a document from April 2010 called "HP StorageWorks Enterprise Virtual Array Compatibility Reference" where the Maximum Virtual Capacity of 300 TB is stated for EVA 8000 and XCS 6.220 .

Thanks again
Best Regards
Horia

[Horia Barbul_2]

I have now another issue with the same EVAs (2C12D+0C5D EVA 8K ) . The customer decided that all disks have to be rearranged so that the DGs have an optimum placement.

The enclosure addresses are non-standard , meaning that there are 8 disk enclosures in the controllers rack (with addresses 1,2,3,4 and 8,9,10,11) and other 9 disk enclosure in the expansion rack ( with the addresses 15,16,17,18,19,20,21,22,23). Does this non standard address allocation changes in any way the disk placement restrictions regarding slots 12,13 and 14 in Disk Enclosure 16,17,19 and 20 ?

Thank you

[Uwe Zessin]

I think the rules still apply due to the way AL_PAs are assigned to specific enclosure/bay positions.

[Horia Barbul_2]

Uwe , thank you for your answer.

That was my guess too but the strange thing is that the customer allready has disks in restricted disk bays : first EVA encl.16 (bays 12 , 13), encl.17 (bay 12), encl. 19 (bay 12) and encl. 20 ( bays 12, 13 and 14). The second EVA has a similar situation. Both storages are - apparently - working fine.

Disks in enclosures 16, 17 and 19 are on loops 1B and 1A while disks on enclosure 20 are on loops 2B and 2A.

The AL_PA addresses for some disks in restricted slots:

enclosure 20 slot 12 04
enclosure 20 slot 13 D2
enclosure 20 slot 14 B9

enclosure 19 slot 12 D4
enclosure 17 slot 12 04
enclosure 16 slot 12 A9
enclosure 16 slot 13 A7

Of course I will rearrange the disks so that the restrictions be respected but I wonder how can this work ?

Thank you

[Uwe Zessin]

Fibre Channel Arbitrated Loop rules:
If a device can't get a 'hard' address, it is free to dynamically allocate a currently unused address (soft addressing).
And it seems that the controller firmware (at least sometimes) tolerates this.