Mark9898
Contributor
Contributor

ESXi 4.0 - configuring storage - I've made a mistake

Jump to solution

I've made a mistake; but I'm too much of a novice to understand where.  Can I please get some direction?

I've built a ESX4.0 box on some old hardware, it has quite large disks, the array ended up being 2.7 TB big, yet the datastore on the end result, are only 741 GB big.  Having a look at the output of fdisk and partedUtil - the entire disk seems to be used (see below).

Can I make use of other 2 TB of storage somehow? Where have I gone wrong?

Also; The how VMFS and datastores layer of disk partitions is confusing to me? Is it disk partion > VMFS > datastore?

~ # partedUtil get /vmfs/devices/disks/mpx.vmhba1\:C0\:T0\:L0

2861100 64 32 1564565504

4 32 8191 4 128

1 8192 1843199 5 0

5 8224 520191 6 0

6 520224 1032191 6 0

7 1032224 1257471 252 0

8 1257504 1843199 6 0

2 1843200 10229759 6 0

3 10229760 1564565503 251 0

~ # fdisk -l

Disk /dev/disks/mpx.vmhba1:C0:T0:L0: 3000.0 GB, 3000080793600 bytes

64 heads, 32 sectors/track, 2861100 cylinders

Units = cylinders of 2048 * 512 = 1048576 bytes

                          Device Boot      Start         End      Blocks  Id System

/dev/disks/mpx.vmhba1:C0:T0:L0p1             5       900    917504    5  Extended

/dev/disks/mpx.vmhba1:C0:T0:L0p2           901      4995   4193280    6  FAT16

/dev/disks/mpx.vmhba1:C0:T0:L0p3          4996    763948 777167872   fb  VMFS

/dev/disks/mpx.vmhba1:C0:T0:L0p4   *         1         4      4080    4  FAT16 <32M

/dev/disks/mpx.vmhba1:C0:T0:L0p5             5       254    255984    6  FAT16

/dev/disks/mpx.vmhba1:C0:T0:L0p6           255       504    255984    6  FAT16

/dev/disks/mpx.vmhba1:C0:T0:L0p7           505       614    112624   fc  VMKcore

/dev/disks/mpx.vmhba1:C0:T0:L0p8           615       900    292848    6  FAT16

0 Kudos
1 Solution

Accepted Solutions
Josh26
Virtuoso
Virtuoso

You cannot utilise a single LUN > (2TB -512b) on ESXi 4.0.

First question would be why you're building such an old version? I'm aware the answer could well be related to support for your older hardware.

The general rule is "one disk/LUN = one partition = 1 vmfs datastore". If you want to break this storage up into two datastores that come in under the 2TB limit, you need to look at what your storage vendor offers around creating two smaller LUNs. I know with, for example, HP hardware, you can create one physical RAID array, with two "local disks" under it to achieve this effect.

View solution in original post

0 Kudos
4 Replies
Josh26
Virtuoso
Virtuoso

You cannot utilise a single LUN > (2TB -512b) on ESXi 4.0.

First question would be why you're building such an old version? I'm aware the answer could well be related to support for your older hardware.

The general rule is "one disk/LUN = one partition = 1 vmfs datastore". If you want to break this storage up into two datastores that come in under the 2TB limit, you need to look at what your storage vendor offers around creating two smaller LUNs. I know with, for example, HP hardware, you can create one physical RAID array, with two "local disks" under it to achieve this effect.

0 Kudos
Mark9898
Contributor
Contributor

Hi Josh26,

Yes, the hardware is only on the HCL for the 4.0 version of ESX.

So as long as I had the logical disks under 2 TB, it'd behave normally, for example I'd be able to have 2x 1.8 TB datastores? But I have to do that a layer below ESX, correct?

0 Kudos
TedH256
Expert
Expert

yes, you do this at the hardware/physical/raid controller level - group the disks into an array, and then (in hP lingo) carve the array into two logical drives each under 2TB. Then esx will see it and be able to create a partition on it, and format it with vmfs (ie create a datastore)

Mark9898
Contributor
Contributor

Thanks all!

0 Kudos