Silly Mistake Adding Storage

[AnthonyM]

Hi,

I have a single ESX server, which up until yesterday had 2 72GB SAS disks with the installation, and VMFS.

Yesterday I added another two 147GB SAS disks, and instead of adding a new logical drive, I extended the drive so that it is 147GB in size.

A reboot later fdisk sees the new size, but I do not see how to use any of it.

The server is a DL360 G5, but appears to be a model were only 4 drives are active, the ACU is installed.

\[ root]# df -h

Filesystem Size Used Avail Use% Mounted on

/dev/cciss/c0d0p2 4.9G 1.4G 3.3G 30% /

/dev/cciss/c0d0p1 99M 29M 65M 31% /boot

none 131M 0 131M 0% /dev/shm

/dev/cciss/c0d0p6 2.0G 54M 1.8G 3% /var/log

\[ root]# fdisk /dev/cciss/c0d0

The number of cylinders for this disk is set to 17840.

There is nothing wrong with that, but this is larger than 1024,

and could in certain setups cause problems with:

1) software that runs at boot time (e.g., old versions of LILO)

2) booting and partitioning software from other OSs

(e.g., DOS FDISK, OS/2 FDISK)

Command (m for help): p

Disk /dev/cciss/c0d0: 146.7 GB, 146745262080 bytes

255 heads, 63 sectors/track, 17840 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System

/dev/cciss/c0d0p1 * 1 13 104391 83 Linux

/dev/cciss/c0d0p2 14 650 5116702+ 83 Linux

/dev/cciss/c0d0p3 651 8584 63729855 fb Unknown

/dev/cciss/c0d0p4 8585 8920 2698920 f Win95 Ext'd (LBA)

/dev/cciss/c0d0p5 8585 8653 554211 82 Linux swap

/dev/cciss/c0d0p6 8654 8907 2040223+ 83 Linux

/dev/cciss/c0d0p7 8908 8920 104391 fc Unknown

Command (m for help): v

143300048 unallocated sectors

What options do I have now?

Thanks

Anthony

[davidbarclay]

AFAIK, the only way to reclaim those disks will be to break the array, which means goodbye install and VMFS.

The install part isn't hard. Re-installation takes about 30 mins, it's the VMFS you need to worry about. Many VMs? Much data?

You could backup the VMs (vmdk, vmx etc) using WinSCP (or Veeam, or whatever). then reinstall, then reconfigure and then restore.

Are you using esxRanger? That would make like a little easier for the backup and restore part.

Dave

[AnthonyM]

Thanks for that.

There are 5 VMs, and most of the data is on different file systems, so it would not be a big issue, but time consuming.

[swamy]

you can take backup of all those VMS using the following scripts From ESXbox

vmsnap.pl

and

vmsnap_all

just run #man vmsnap.pl and #man vmsnap_all

this thing is on ESX 2.X

on ESX 3.X vmsnap_all has been deprecated with this release of ESX.

Please use vcbSnapAll instead.

for more information u can try the following vcbmounter and vcbrestore

good luck

[VirtualKenneth]

Alternately you can create a NFS mapping to your ESX host, format that as VMFS and move the systems to the NFS location temporary.

[AnthonyM]

Ok, everyone was correct that the only way to recover the lost space was to break the array.

For anyone in future, here is what I did.

Replace the first 72GB drive with a 147GB, rebuild the array.

Replace the second 72GB drive in the same way.

Put the 72 GB drives on another controller, creating a new array. Recreate the partition structure. Shutdown all the VMs, and use DD to copy the partitions, one at a time exactly.

Used the ESA install disk to to a console, and manually re-installed GRUB on this new array.

Rebooted, using the new array.

Fixed the fstab entry for the swap, and activated it. Ensured the VMs were working.

Broke and re-created the array on the 147 GB drives to gibe me all the space back.

Quite a long process, and a fw hiccups, but worked!