I live in fear of APD situaltions, and I need to remove/delete several iSCSI datastores. What is the process to follow for this, given vSPhere 5.0?
1)remove all virtual disks
2)is there an un-mount process that I must do from each host? I think in 5.1 "unmount" is in the Vsphere client GUI - how do I do it in 5.0?
3)Delete datastore from one host
3)rescan for datastores from other hosts
Is this process correct? Anything else I need to do to ensure that I don't end up with an APD?
Hi,
You can perform this process via the GUI under Host > Configuration > Storage > right click on VMFS
Some additional things to consider;
VMware will provide a warning if all the pre-reqs are not meet when trying to remove a VMFS.
Jeremy
I always follow the instructions in http://kb.vmware.com/kb/2004605
I have found this KB:
But per usual, the article makes some vague statements that ... I am not sure what to do with. Specifically in the section below it says : "Caution: You should successfully unmount the datastore using the documented method prior to completing the steps described below" - bu then the steps below are, in fact, the process for unmounting" !?!?!
What the heck - so do I just:
use the vsphere client to:
1)perform the unmount for each host (or can be done for all hosts at once from the datastores and clusters view), then
2)perform the "detach" operation for each host individually from the vsphere client (step 3 below)
3)unpresent the LUN using vendor-specific storage tools/consoles
4)rescan from all hosts using the vsphere client
Is this correct?
To unmount a LUN from an ESXi 5.0 host using the vSphere Client:
When the device is detached, it stays in an unmounted state even if the device is represented (that is, the detached state is persistent). To bring the device back online, the device must be attached."
Brian Atkinson wrote:
I always follow the instructions in http://kb.vmware.com/kb/2004605
yes, but see my last post - I'm a nervous nelly maybe, but I sure hate it when technical articles are written so poorly. Is my short list of steps correct?
yes so some more to chew on. From this part of the documentation:
we learn about unmounting, and from this:
We learn about detaching.
These are seperate things, apparently - but detaching seems to also unmount. So is there any reason why not to detach the datastore one at a time (which also unmounts it) rather than doing unmounting then detaching as seperate steps? Am I missing something here?
If this doesn't help, set claimrules. This is like dropping an anvil on an egg in order to cack it, but in the end it does the job 😉
fr0nk wrote:
If this doesn't help, set claimrules. This is like dropping an anvil on an egg in order to cack it, but in the end it does the job 😉
If what doesn't help - the procedure I sketched out?
I guess I just want to understand what exactly I need to do to avoid an issue - if the "unmount and then detach" process works and is what I am supposed to do, then great! But the docu and kb leave so much to be desired in terms of being clear!