Hi,

I'm trialing RHEV and have a lab that I work on.. I usually use VMware on this lab, but as my VMware setup uses autodeploy, I can install RHEV on the nodes of my cluster and boot from RHEV or change boot order and PXE boot back to ESXi.

So far so good, added my hosts in the cluster, I have an old CX200 direct attach SAN that still works well and has a couple of terabytes of storage. I unmounted one of the LUNS from my ESXi hosts so no hosts are associated with the LUN and the ESXi hosts are now RHEV hosts. I have RHEV-M loaded and added the hosts to my cluster. Now for storage. I go to add a new storage domain but it says the LUN I have designated for RHEV is in use. So I tick the box anyway and tell it to use the LUN, another warning, but I tick the box to say override the command, hit ok and then it fails and says it cannot create the volume group on the LUN. That's it, There are a couple of LUNS that appear as free and I have compared them on the SAN and can see no difference at all between the free LUN and the LUN I am trying to use. I have unbound the LUN and rebound it so any data on it has been destroyed.

So for the moment, my virtual experience with RHEV has stopped.., any RHEV experts out there that have any idea on what is happening?

Thanks

well, as usual, after asking the question, the answer was found of it's own accord. Obviously there is no automatic rescanning of the FC after changes are made. I imagine there is a way to force a rescan, but not from the RHEV-M interface it would appear, so I expect there is some form of command to rescan the SAN from the command line, but as I didn't know now to do that, I tried a reboot of the host and the LUN is now free and I have added my storage domain...

Normally, you'd be able to avoid the entire experience if you presented RHEV with an empty LUN. I usually simply dd if=/dev/zero of=/dev/mapper/MYLUN bs=1024 count=20000 to wipe the beginning, at this point it just looks empty to RHEV and it can create it's own LVM structure on the LUN turning it into a storage domain

the problem is that I have esxi hosts connected to the SAN as well and they are greedy... ie they mount anything and everything as soon as they see it... so have fixed the problem for now by masking the LUN on the esxi hosts...

this is what zoning is for

it's direct attach, masking is my only option, hopefully getting a switch soon so I can zone, but it's just my lab and I have to take what ever I can get. Masking the LUN at the ESXi host works anyway and it was one of the tasks on my VCAP exam that I though I'd never have to use :-)