VMware Cloud Community
js_opdebeeck
Enthusiast
Enthusiast

Urgent - LUN are disapearing on different ESX

Hi ,

It's friday and problems are behind the door.

The different iSCSI storage are like a REMOVE command has been launched ... and the unused Datastore diaspeare ..

I think only running iSCSI connection maintain the current Datastore. If I stop the machines on it I think Datastore will disapeare .

I put a new ESX Test server ... if I connect it to the iSCSI storage, the LUN are 'blank' ... I don't go farther, I don't want to 'format' or erase anything.

Attached .. some screens

Any idea ???

Thanks

0 Kudos
8 Replies
js_opdebeeck
Enthusiast
Enthusiast

More info ...

On different ESX ... the unused/unlocked Storage is already 'removed'.

1. Is it possible to see if a 'REMOVE' command has been launched and revoke it on the current alive Datastores

2. If not possible and if I lose my Datastores Smiley Sad , is it possible to 'remount' it without losing it's content ?

It's really a bad friday ...

More screenshot , more logs ...

0 Kudos
js_opdebeeck
Enthusiast
Enthusiast

Any news ???

I forced a Snapshot on my Equallogic and MD3000i ...

Maybe VMFS Volume Resignaturing can help me .

Any suggestion?

Thanks.

Js

0 Kudos
kjb007
Immortal
Immortal

Resignature would come in to play if your host can see the LUNs. Run esxcfg-rescan vmhba<0|1|2> on your hba's, to see if the iSCSI LUNs are seen. If they are, look in the vmkernel log to see if the LUNs are being seen as snapshot LUNs, and they are being disabled. Before I resignature, I would try and turn off snapshot disable by going into the ESX server advanced configuration, and set the LVM.DisallowSnapshotLun to 0. Then, rescan again, and your vmfs should show up correctly under the /vmfs/volumes.

-KjB

vExpert/VCP/VCAP vmwise.com / @vmwise -KjB
0 Kudos
js_opdebeeck
Enthusiast
Enthusiast

Thanks ...

I'll test that on monday (21st is National Day in Belgium), and I'll give you a feedback.

I've some 'removed' Datastore still mounted, because the connexion is still alive on them (VMServers running), do I have to shut down all of them before ? or can I force a rescan ?

Thanks.

Js

0 Kudos
kjb007
Immortal
Immortal

If you have vm's running, do not shut them down. As long as those vm's are up, the datastore can't be fully gone. If you have mission critical vm's, then you can also use vmware converter with the virtual machine and convert the vm to another datastore that is still good. You can do the rescan either way. I would open a case with vmware, they would be able to help you directly so as to minimize any data loss, if you really have lost datastore.

-KjB

vExpert/VCP/VCAP vmwise.com / @vmwise -KjB
0 Kudos
js_opdebeeck
Enthusiast
Enthusiast

Hi ,

Good news .. Smiley Happy

After a long day on call with DELL support and VMWARE Technician, the issue is solved

It was a problem of VMFS Partition table that was corrupted on the Storage.

Short answered ... playing deeply with FDISK, and some different VMware command-line tools to bring alive the partitions ....

Note that if you use iSCSI storage and you have increased the Datastore size ... you need to know exactly the size of each volume and extends ...

I've the history of the sequence of commands ... I can publish some sequence if you want.

Thanks for help.

Kind regards.

Js

0 Kudos
kjb007
Immortal
Immortal

Glad you got it all working!!! I would love to see that history you saved and the process/commands you followed. You can post them, or message me with that info.

-KjB

vExpert/VCP/VCAP vmwise.com / @vmwise -KjB
0 Kudos
js_opdebeeck
Enthusiast
Enthusiast

Here is a full document with a summary of the different task and commands.

I'm a bit currious ... where to find information about "How/where VMWARE ESX remove DATASTORE".

Js

0 Kudos