VCSA 6.7 doesn't show correct datastore size

[hoFFy84]

Hello,

I have a cluster with two hosts running on 6.7.0, 10302608. I'm using hyper converged SAN Symphony from Datacore to replicate the storage between these hosts. Both hosts have had a RAID with a size of 16TB. These 16TB are presented to the Datacore VMs which replicate it and serve it as a shared storage to both hosts.

Then we created another RAID in the hosts, and presented that new datastore to the Datacore VMs. In Datacore, we expanded the LUN to 20TB. In the vCenter we then expanded the LUN presented by Datacore also to 20TB.

At the vCenter, when choosing a host -> Configure (my be different, I'm using a german system) -> Storage Adapters, I see a capacity of 20TB on both hosts:

When I then click on the Datastore itself, It shows only a capacity of 16TB:

But if I click on "Geräte Backing" (might be "device backing"?!) I see 20TB again:

I also can't expand any VMDK with disk space that would use more space in total than these old 16TB. All hosts are running the same vSphere Version and i rescanned several times the HBAs..

What else can be done?

[a_p_]

In Datacore, we expanded the LUN to 20TB.

In the vCenter we then expanded the LUN presented by Datacore also to 20TB.

Let's make sure that I understand this correctly.

First you expanded the LUN on the storage side, then you did increase the VMFS datastore size with the additional disk space (16TB -> 20TB). Is that what you mean with your second statement?

In this case, connect to each of the ESXi hosts directly, and rescan the storage to see whether the new size shows up. Once both show the new size, you should then see the new size in vCenter Server too.

André

[hoFFy84]

Hi,

yes you unterstood correctly. I tried your suggestions but that didn't help.

This what I see on both hosts:

Under datastores I see 16 TB (and I'm no longer able to expand the disk)

but under devices I see 20TB

[a_p_]

Please follow steps 2 to 7 from the Resolution section in https://kb.vmware.com/kb/2002461​, and post the results for steps 5, and 7.

For step 2, enable the SSH service on tan ESXi host, and connect to it using e.g. putty.

For step 5, please run partedUtil getptbl ... (instead of get).

André

[hoFFy84]

Hello,

the output from 5 is:

[root@WMVM4:~] partedUtil getptbl "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833:1"

unknown

2734847 255 63 43935332319

but that differs from the example in the KB, so I repeated step 3:

[root@WMVM4:~] vmkfstools -P "/vmfs/volumes/S01_L00_20000/"

VMFS-6.81 (Raw Major Version: 24) file system spanning 1 partitions.

File system label (if any): S01_L00_20000

Mode: public

Capacity 17592186044416 (16777216 file blocks * 1048576), 3881018851328 (3701228 blocks) avail, max supported file size 70368744177664

Disk Block Size: 512/512/0

UUID: 5a952cee-12c46082-6207-901b0eff07f7

Partitions spanned (on "lvm"):

        naa.60030d90d088500635dfd3a3c3f16833:1

Is Native Snapshot Capable: NO

but If I understood correct, then "naa.60030d90d088500635dfd3a3c3f16833:1" is the correct device identifier, isn't it?

Which commando should I use at step 7?

If I run the following I get an error:

[root@WMVM4:~] partedUtil getUsableSectors "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833:1"

Unknown partition table on disk /vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833:1

[a_p_]

Devices and partition look very similar.

You'll need to remove the partition number ":1" in this case to get the expected results.

partedUtil getptbl "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833"

partedUtil getUsableSectors "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833"

André

[hoFFy84]

Hello André,

the output is as follows:

[root@WMVM4:~] vmkfstools -P "/vmfs/volumes/S01_L00_20000"

VMFS-6.81 (Raw Major Version: 24) file system spanning 1 partitions.

File system label (if any): S01_L00_20000

Mode: public

Capacity 17592186044416 (16777216 file blocks * 1048576), 3881005219840 (3701215 blocks) avail, max supported file size 70368744177664

Disk Block Size: 512/512/0

UUID: 5a952cee-12c46082-6207-901b0eff07f7

Partitions spanned (on "lvm"):

        naa.60030d90d088500635dfd3a3c3f16833:1

Is Native Snapshot Capable: NO

[root@WMVM4:~] partedUtil getptbl "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833"

gpt

2734848 255 63 43935334400

1 2048 43935334366 AA31E02A400F11DB9590000C2911D1B8 vmfs 0

[root@WMVM4:~] partedUtil getUsableSectors "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833"

34 43935334366

[a_p_]

Not sure what happened, but it looks like the resizing didn't run to the end.

The partition has been resized, but not the file system.

The below steps should work (see also the KB article I posted earlier):

1. run a rescan
   vmkfstools -V
2. let partedUtil fix the GPT if required
   partedUtil fixGpt "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833"
3. resize the VMFS file system in Partition 1
   vmkfstools --growfs "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833:1" "/vmfs/devices/disks/naa.60030d90d088500635dfd3a3c3f16833:1"
4. do another rescan
   vmkfstools -V
5. check whether the expected capacity shows up as now
   vmkfstools -P "/vmfs/volumes/S01_L00_20000"

The rescan, and capacity check may also be done from the GUI, as you will have to rescan the other host too, to make it aware of the new datastore size.

Since we are dealing with a huge capacity, and data, please consider to open a support call with VMware to verify the steps.

André

[hoFFy84]

Hello André,

this has been resolved in a WebEx Session together with the support.

He also tried your step 3 and after several rescans we saw the correct size.

Thank you!