keep the existing signature vs assign a new signature - which impact?

[bonne1]

After upgrading our SAN, our iSCSI targets automatically got a new name. We were still able to see the SCSI targets, but we could not access the datastores untill we clicked "Add storage" under "Configuration", "Storage". Then the datastores showed up on a list, and when we added each one of them, we were promted to choose from the following selections:

1) keep the existing signature

2) assign a new signature

3) format target

As we would like to keep our data on the stores, we were sure about not selecting format (-;

We expected that selecting to keep the existing signature, data on the disk would remain untouch, which is why we selected this. We had to do this on all our ESX servers which has access to this store. Well, everything is up and running again.

Now, we are unsure as to which impact this will have on reboots as well as when giving future ESX servers access to the SAN. Can some of you enlighten us on what exactly is the issue here as we are really not sure of what is going on?

Thanks in advance, Lars.

[bonne1]

After a reboot of one of the servers, we are back again to the point where we cannot access the datastores (they are not listed even after a refresh of the datastores).

In order for them to keep comming "online" after future reboots, I can then imagine, that we would need to do an "assign a new signature" rather than keep. Or?

If correctly assumed, which impact will this get on the data of the datastores - will it be kept untouched after assigning a new signiture so that data prior to assigning new signature will be able to be accessed again?

Regards, Lars.

[SimonStrutt]

I would suspect that when your ESX's are rebooted they are looking for the old iSCSI target name and so can't the datastore's until manually pointed to them by you.

The only option that is destructive to your VM's on the datastore is the "Format Target" option, therefore it would be logical to try assigning a new signature to the affected datastore(s), and refresh the storage on your other ESX's.