VMware Cloud Community
RTX
Contributor
Contributor
‎07-08-2009 09:42 PM

Partitions lost

This night all our file servers has lost their extra disks (D:\ E:\ etc.). The disks has RAW mappings to our SAN, we can see the LUN's in the storage adapters, but the file server cannot recognize the disks.(see dump). All these disks were GPT-disks. All the other servers we have do still have their RAW mappings. It seems that its only our GPT-disks that are gone.

Yesterday we upgraded two of the ESX servers from ESX 3.5 Update2 to ESX 3.5 Update 4, but nothing seemed wrong then.

Any help is appreciated as we have no access to our four file servers data.

Preview file
71 KB
0 Kudos
3 Replies
depping
Leadership
Leadership
‎07-08-2009 10:51 PM

did you upgrade the hosts with all the LUNs still attached to it? (did you pull the fc cables or unpresented the LUNs?) If not, I would recommend not touching the server anymore and calling support. They can probably still help you and recover your data.

Duncan

VMware Communities User Moderator | VCP | VCDX

-

Blogging:

Twitter:

If you find this information useful, please award points for "correct" or "helpful".

0 Kudos
RTX
Contributor
Contributor
‎07-08-2009 11:02 PM

Unfortunately yes - nothing was unattached, we have done this xx times without problems.

Really hope VM Support can do something, I have 130 engineers doing nothing and missing 3 TB data.

Tried a work-around with Fdisk, I found here in the VM forum, but without success, now restoring data for that partition (only 240 GB though) Smiley Sad

Thanks for your reply

0 Kudos
RTX
Contributor
Contributor
‎07-10-2009 04:49 AM

Well, our VMware partner gave up on this yesterday, so the last 36 hours I have been restoring 3 TB data back to new LUN's. At least now I know our backup-strategy and plan works Smiley Happy

I still have 3 esx-servers to upgrade and I have never heard this about unpresenting the LUN's or pulling out the fibercables to the server being upgraded. Do you or anyone else for that matter have a whitepaper on an upgrade procedure. I have been reading this Upgrade 3.5 u2 but did not find anything on this.

What I'm looking for now is a procedure that ensures that this does not happen again on the next upgrade. (having 130 Software/ Hardware engineers doing nothing for 1 day is a very costly affair.....)

0 Kudos