Losing LUNS from SAN Storage in VMware ESX 3.0.1

[deepvrce]

hi,

I have VMware ESX Server 3.0.1 installed on a SunFire V20z connected to SAN Storage (Sun Storedge 3510). VMware to start off is able to see the LUNs properly and I could use them too with the Guest OSes. The problem I am facing is that if I reboot the VMware ESX Server, then it starts reporting 0.0 GB as the capacity for the LUN. The only way to get around this seems to be unmapping the LUN from the Storage Box and then mapping it back again, but with a different partition on a Logical Drive on the Storage Box. At least I have not been able to get around this issue yet. Has anyone seen anything like this before?

And I have 2 such server nodes and faced the problem with different LUN/partition combinations on both of the nodes. Does VMware cache or store SAN specific information which needs to be cleared? The node doesn't work with the same partition on the Storage Box once it reports 0.0 GB the first time.

Any help would be great..

[VirtualKenneth]

Well I'm not sure if this could have anything to do with your problem but the V20z isn't on the compatible list http://www.vmware.com/pdf/vi3_systems_guide.pdf

When you presented the LUN and created the host with the SAN management software, did you define that the host is a VMware ESX system rather then a default "Windows" system?

[deepvrce]

What difference would it make to define the host as a "VMware ESX" System rather a Windows system? Not that I know how to do that, or if at all if can be done with the Storage Box that I am using (Sun StorEdge 1500).

[wobbly1]

it sets connect specific strings for the operating system to read the lun information correctly. how is your zoning configured?

[deepvrce]

I have not configured zoning. LUN mapping from the Storage Box is what decides what the nodes see.

[wobbly1]

that may not be helping. I would suggest zoning each server hba into a zone with access to the storage controller(s) as a first step along with setting the connection type on the storage to ESX server if the option is there

[deepvrce]

ya that definitely makes sense to me.. But what is puzzling is that things were working fine.. I rebooted the ESX Server and then after coming back it started showing 0 GB size for the LUNs which were mapped earlier. Does the server cache LUN and partition information anywhere? Coz the only to get things working was to use a different partition on the SAN box.

[wobbly1]

there is an element of caching going on which is why you have to rescan after certain activities however zoning issues can throw up all sorts of strange problems and erratic behaviour and the same is likely with having the wrong connection type/string set for the individual luns

[deepvrce]

thanks.. will try that out sometime.. hopefully i would be able to get around this strange problem.