EVA6000 - Recover space from deleted VDisks

[Rob Costello msft]

Hi all.

I'm very new to HP Storageworks gear, and I have a problem.

I have an EVA6000, which had several vdisks in place. Since re-organising some servers and needing to change disk config, I've deleted the vdisks in Command View EVA, but the disk space that those disks had is not freeing up.

For example, I have 100Gb free space, and I delete a 200Gb vdisk, but I can still only see 100Gb of free space. I can't create a vdisk greater than 100Gb.

Can someone tell me how to recover the space from these deleted vdisks please?

Thanks,
Rob Costello.

[Mark...]

Hi,
Hmmm, interesting - all you should need to do is to delete the vdisks and that's it.
Have you got more than one disk group??
If you have then when you try to create your new vdisk then use the dg drop down to select the dg with the space in it - ie the dg that probably had the vdisks in it that you have deleted.
If not that then what version of XCS and cmd vw are you using ??? I am not aware of any problems like this but maybe others have seen this?
You could also go to the left hand pain of cmd vw, select the top item and then do "rediscover". Basically this gives the fabric a "kicking" to wake things up abit. Check space again.
Do you have any warning triangles within cmd vw ?? if so investigate.
Do you know what the sparing was before you deleted the vdisks ?? The sparing requested level may have been requested higher but not implemented because of lack of space. Once you deleted the vdisks then space would have been created and the sparing level would have adjusted itself. This could well have taken all your "free" space!
To resolve, adjust your sparing level as required.
If non of the above works/helps then you could go for the - kill / logoff cmd vw and then go back in again and check things out.
Update as to how you get on.
Mark...

[Rob Costello msft]

Hi Mark,
There is only one disk group setup.
I'm running version 6.0 of Command View (that's the latest from what I can tell).
I tried running the Discover, as well as a Refresh of the array, but still nothing.

I've got one triangle which is for "Unmappable hardware", not sure what that is.

It was set to Double for the sparing level, and that gave me a triangle too. Although it was set to double, it was saying the actual level was Single. I've dropped the requested level to Single and the error triangle is gone on that node.

I've tried restarting CmdVw but that doesn't seem to do much.

Thanks for the help so far, any other things I can try?

Thanks,
Rob

[Mark...]

Hi,
Yes, warning triangles are a problem!!
If you have "unmappable h/w" with a warning triangle over it then you do have a problem and this may well be the reason for the loss of disk capacity.
Unmappable h/w is typically a problem with the enclosure address bus , but there are several things to check. The EAB is made up of a seriess of cables that run down the left hand side of the back of the enclosure. Some of the connectors then have cables running to the bottom port of each shelf EMU and a "Y" cable to the ctrl pair. The EAB should also be terminated at both ends. You should check to see that each EMU has a number on the 2 digit display - this is the shelf ID. It should tie in with what you see in cmd vw. Open up the unmappable object in cmd vw and see what get. At a guess it will possible be a whole shelf, investigate as required.
In a nutshell you need to check out all Eab cabling / termination & EMU's.
If you do find any problems then once it is fixed unmappable h/w errors can take quite a while to disappear/sort themselves out in cmd vw, I don't mean hours but maybe up to 30mins or more.
If you have a h/w contract with HP or other vendor then you may wish to get them involved?
Mark...

[Mark...]

Hi,
Had this HP advisory the other day. It may help solve your problem if you have not solved it already. Worth a read anyway:

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c01035871&dimid=1256978157&dicid=alr_jul07&jumpid=em_alerts/us/jul07/all/xbu/emailsubid/mrm/mcc/loc/rbu_category-Advisory(Revised)/alerts

Mark...