Newly formatted 1 TB+ shared VMFS datastore went missing after few days of operation.

[albertwt]

Hi All,

 

I’ve just reformatted 1.44 TB LUN from my Dell MD3000i iSCSI SAN this was a Linux Ext3/FS RDM attached into one of my VM in on ESXi02 host, i did the following:

1. detaching the RDM from the VM

2. formatting it from the Dell SAN management software

3. present it to both host of ESXi01 and ESXi02 so that this new LUN can be shared together as VMFS.

4. Go to any of the ESXi01 or 02 settings and then add this datastore, format it with 8 MB block size and then make it available to both hosts by creating proper naming.

 

Today suddenly it went missing without no apparent reason or error? I’ve patch both host few days ago and restart both host and the new LUN was still ok yesterday, but today it went missing so I had to reformat it again from the VCenter from Datastores view.

 

I tried:

1. Rescan datastore – nothing appears

2. Refresh the datastore view from both hosts - nothing appears as well ?

 

I wonder what was happening before and I couldn’t imagine if that new LUN got data in it ? i might be in deep trouble if so.

 

Any kind of comment and suggestion would be greatly appreciated.

 

Thanks.

 

Note:

Vcenter Server 4.0 208111

VMware ESXi 4.0 244038

[Duke]

I assume if you go onto the MD3000i then the LUN still exists and has the correct details, iSCSI initiator security, etc? Can you add the iSCSI initiator of a Windows box then use the MS iSCSI tool to try to connect to the LUN and see if there's any data in it? vCenter should have logs - any info in them?

[albertwt]

Thanks for the reply Duke,

I didn't try to mount that RDM into the Windows VM or box as I've just reformatted that partition with VMFS 3.33

 

I don't know why suddenly it went missing ?

[Duke]

Not a clue to be honest, I was just wondering if a completely separate Windows box would even see the LUN on the SAN is all. Does the Dell management software say the LUN still exists?

[albertwt]

yes it is still exist, i was suspecting around these area:

 

1. After ESXi 4 patch, i did the restart for all of the host one by one. (if it is totally new install i can understand that the iSCSI connection string must be changed / remapped again).

2. After the VCenter Server restart due to the Windows update.

 

can any of those becoming the source of the problem ? I'm worried if something does happens again during the production cutover then it will be a disastrous.