Datastore displays incorrect Capacity

[ImedecsIT]

Hello,

We recently installed ESXi onto a whitebox we had laying around to test it out. The machine initially only had a 40 GB iDE drive in it. We wanted to try a larger VM so we needed more space so we added a 20 GB ide drive to test it out.

It turns out that after we added that 20 GB drive to the machine and added it to the datastore it began to fail. We hadn't added any more information to the ESXi (ie more VM's) so there wasn't anything on the 20 GB drive. So we powered down the host and removed the drive. When we powered back up, the Datastore still displays the capacity with both drives installed.

Is there any way to get the datastore to correctly display the capacity? I did try via VI Client the configuration tab-storage adapters-rescan and that didn't work. I also tried to refresh on the Storage view but no luck there either. Please advise as I have a new 250 GB drive I'd like to add to do mroe testing but I want to hold off until this is resolved.

Regards,

Brian

[vheni79]

Hello,

I am not shure if this will help.

Login to VI -> register config -> storage -> click on rescan.

Regards,

Nico

[ImedecsIT]

Nico,

I see the Configuration tab in VI and I have gone to Storage and tried refresh, I do not see a rescan. Is register Config something different from the Configuration tab?

~ Brian

[mike_laspina]

Hi,

Did you extend the vmfs volume with the 20GB drive?

[vheni79]

Brain,

sorry for my pore english.

I mean "rescan" at the storage section.

Nico

[ImedecsIT]

Nico, no worries, I thought thats what you meant.

Mike,

I did extend the vmfs with the 20 GB drive by right clicking on datastore and doing properties. I then selected the drive and Added Extent. Does that mean you have a solution. I am reading your blog and it is very complex but I think i can probably muck my way through it. please let me know what steps to follow.

~ brian

[mike_laspina]

There is no GUI based automated way to correct the extent definition within the vmfs volume. It would require vmfs header surgery to remove the extent definition.

I have not examined this senario before, regardless of that its not an easy fix.

If you are interested in pursuing the surgical removal of the definition you could post the hex output and I can examine it. You would need a hexeditor to fix the issue.

Regards,

Mike

[ImedecsIT]

Mike,

Thanks for the info. Not sure i'm up for a surgical removal considering this is a test box. Is this something that is common to occur if there is a single drive failure?

Is there some way for me to remove the iso image and the 1 configured VM i have on the box so i can just reinstall the whole thing onto my 250 GB drive?

~ Brian

[mike_laspina]

You should be able to use the browse datastore interface to transfer the vmdk's and ISO file to a workstation. You can then rebuild the ESXi host with the 250.

Then you can transfer the files back and import the VM.

If you have extents and a volume fails you will lose any VM that spans the blocks of the two devices and all VM's on the impacted extent, not recommeded with raid 0 in my books.