VMware Cloud Community
dm71
Contributor
Contributor

Lost a datastore

I have a small cluster of ESX 3.5 hosts connected to a san. There are 3 datastores, each on its own lun. The third datastore (we'll call ds3) had two extents (the array was extended at one point), first extent was about 1TB and the other was about 400GB. Today ds3 suddenly went away - meaning - it was still listed in the gui, but when I go to browse it, there is nothing in it. And now I see that the datastore properties shows the first extent is 1.4TB and the other is 0. And it no longer thinks its VMFS, instead it says the primary partition is HPFS/NTFS. All nodes see this same corrupt configuration.

Why did this spontaneously happen? Is it considered normal for a datastore to just go awol? And of course, how do I get it back? I can't manually reconfigure the device with fdisk because I have no way of knowing what the true extent sizes were. It seems that that's where I went wrong when I grew the array over a year ago. Is that not supported?

I sure would appreciate any help in this matter. Thank you.

Reply
0 Kudos
4 Replies
lamw
Community Manager
Community Manager

Did you try a esxcfg-rescan <vmhba_adapter> on one of the ESX Host? If there are running VMs you can also put one of the hosts in Maintenance mode and restart the system to see if it catches the storage. We recently added new LUNs to our cluster and there were few offending systems that acted up, we resorted to rebooting the host after restarting the agents and re-scanning for the newly added LUNS. Hopefully your extent was not corrupted and just small glitch from VC not displaying correctly, as you know when you extend a VMFS volume, the starting volume has all the meta data for your second VMFS volume you've combined, so if you corrupt the primary, you can technically lose all data to the extent. Hopefully this is not the case

Reply
0 Kudos
joergriether
Hot Shot
Hot Shot

Did you use VCB on this partition recently? If yes, could it be, you forgot to disable automount feature? Check out this one here http://www.riether.com/2008/06/vcb-proxy-forgot-to-disable-automount.html if it fits but remember you have to call vmware this is no fun and should be adressed asap by vmware itsself they know how to fix these issues.

best regards

Joerg

Reply
0 Kudos
dm71
Contributor
Contributor

Thanks for the replies guys. It turns out that a co-worker killed it. He accidently mapped his windows box to this lun instead of the new one he created. Everything is gone. I have to go to tape now.

Kev

Reply
0 Kudos
lamw
Community Manager
Community Manager

Hm ... he was able to map his Windows system to a VMFS volume? Usually the case is through an RDM and if it's set to physical using SAN utilities, you could access the LUN versus virtual where you still have access to use snapshots.

Reply
0 Kudos