Damn ! I get “Unable to Read Partition Information”

You rescan your HBAs on the ESX and your new LUN suddenly appears. But when you want to add it as new datastore you get “Unable to Read Partition Information” . The same error happens on every ESX you try.

Ok, you have to know that only MS-DOS partitioned disks are recongnized by ESX. I had to face this issue once Storage team gave me the disk. It was re-assigned from a previous AIX system. So , heres what you can do :

Firts: Ensure the LunID and path are correct because all data will be destroyed.

1. esxcfg-vmhbdevs -q | grep LUNID

This will show you the canonical path for your LUN and the devicename

Use the devicename from step 1 to invoke parted tool.

2. parted /dev/sdc (The path showed before)

A new prompt will appear waiting for commands

3. mklabel msdos
4. Quit and try to Add a new Datastore again.

Good Luck !

VMware KB

Converting to Thin Provisioning

There will be a moment in your life when your DataStore will run out of space or your ESX 3.5 Virtual Machine should be exported to a new brand ESX 4.0 server and you shall decide to take advantage of Thin Provisioning. That day you’ll need to deal with ‘vmkfstools’.

Basically the process is rather simple but your VM must be powered off and must have the same free space in a datastore as your current vmdk file size. Then just connect to your ESX console – putty’s time – become su and run this command:

vmkfstools -i /vmfs/volumes/VMname/VMname.vmdk /vmfs/volumes/VMname/VMname-thin.vmdk -d ‘thin’ -a lsilogic

The cloning process will start and you’ll get a thin copy of your thick vmdk file. Let’s edit your VM from VI client to remove -without deleting- the old disk and add your thin one as an existing disk. Check that your VM starts and runs properly. Then shut it down and browse your VM datastore, check how much storage is really using your new thin vmdk file. Once recovered from the shock 🙂 delete your old vmdk file.