VMware Cloud Community
aughsydney
Contributor
Contributor
Jump to solution

VMWARE 6.7 MSA2040 and Increasing Datastore size

Hi There,

I need to know how you go about increasing the size of a datastore in 6.7

In my case I logon to the MSA and increase the size of the Volume (the DataStore volume), however, in the vSphere web client the datastore size does not change and I have to use the following series of commands before the datastore displays the correct value.

1. On the MSA increase the size of the Volume
2. Using SSH Logon on to the esxi host and run these commands:
a. partedUtil fixGpt "/vmfs/devices/disks/naa.600c0ff00029475d0e5a605b01000000"
b. partedUtil getUsableSectors /vmfs/devices/disks/naa.600c0ff00029475d0e5a605b01000000

Make a note of the large number
c. /etc/init.d/storageRM stop
Stops the storage monitor (has no affect on the running esxi system)
d. partedUtil setptbl /vmfs/devices/disks/naa.600c0ff00029475d0e5a605b01000000 gpt "1 2048 2539061214 AA31E02A400F11DB9590000C2911D1B8 0"
The values above are got from running this command :
offset="128 2048"; for dev in `esxcfg-scsidevs -l | grep "Console Device:" | awk {'print $3'}`; do disk=$dev; echo $disk; partedUtil getptbl $disk; { for i in `echo $offset`; do echo "Checking offset found at $i:"; hexdump
-n4 -s $((0x100000+(512*$i))) $disk; hexdump -n4 -s $((0x1300000+(512*$i))) $disk; hexdump -C -n 128 -s $((0x130001d + (512*$i))) $disk; done; } | grep -B 1 -A 5 d00d; echo "---------------------"; done


This command returns the information above – find the datastore you are extending and make a note of the line above.
In the example give the value 2539061214, was obtained by running the command at 2b.
partedUtil getUsableSectors /vmfs/devices/disks/naa.600c0ff00029475d0e5a605b01000000

e. Now run this command :
vmkfstools --growfs "/vmfs/devices/disks/naa.600c0ff00029475d0e5a605b01000000:1" "/vmfs/devices/disks/naa.600c0ff00029475d0e5a605b01000000:1"
  Check the vsphere web client you should find that the DataStore is now the correct size.
f. Restart the storage monitor
/etc/init.d/storageRM start

Like really all the palaver to increase a datastore!! (Please don't nit-pick the commands above - it might not be exactly correct - but its roughly what I had to do.)

In my previous position I worked in a Netapp environment - when I wanted to increase the size of the datastore - all I had to do was ask the storage guys to increase the volume/LUN on the Netapp Filer and bang! my datastore displayed the new size (without the need to perform all these esxi steps)

Is this a limitation of the MSA 2040?  You have to issue commands on the vSphere side before you can see the new datastore size?

ta

Reply
0 Kudos
1 Solution

Accepted Solutions
IRIX201110141
Champion
Champion
Jump to solution

Iam pretty sure that your Storage Admin present the Volums as a "File" storage trough NFS and than you have nothing to do on the Host site.

If you deal with Block storage you have to do

1. Increase the LUN on the Storage

2. "Rescan" the HBA on your first Host

3. Select the Datastore and  choose Expand

4. Perform a Rescan on the Cluster if there is one so the other Hosts take notice

All is possible within the GUI and doesnt takes longer than 2minutes if you do it the first time.

If your Storage vendor provides a plugin... there is a integration into the WebClient and you can just select expand Datastore within that plugin.  I cant speak for the MSA2040 and also support for "older" models lack of software support for vSphere 6.7 and the new HTML5 Client in general. But if its available all of the steps from above are orchestrated by the plugin.

Regards
Joerg

View solution in original post

Reply
0 Kudos
6 Replies
dbalcaraz
Expert
Expert
Jump to solution

Did you try to rescan storage adapters after the resize?

-------------------------------------------------------- "I greet each challenge with expectation"
Reply
0 Kudos
aughsydney
Contributor
Contributor
Jump to solution

Yes - I rescanned - eventually I also rebooted one of the esxi hosts - no joy - the datastore size remained the same.

From your experience could you tell me if I'm correct - if I increase the volume size on the MSA, there is no need to do anything on the esxi side? - the datastore should show the increased size?

Reply
0 Kudos
IRIX201110141
Champion
Champion
Jump to solution

Iam pretty sure that your Storage Admin present the Volums as a "File" storage trough NFS and than you have nothing to do on the Host site.

If you deal with Block storage you have to do

1. Increase the LUN on the Storage

2. "Rescan" the HBA on your first Host

3. Select the Datastore and  choose Expand

4. Perform a Rescan on the Cluster if there is one so the other Hosts take notice

All is possible within the GUI and doesnt takes longer than 2minutes if you do it the first time.

If your Storage vendor provides a plugin... there is a integration into the WebClient and you can just select expand Datastore within that plugin.  I cant speak for the MSA2040 and also support for "older" models lack of software support for vSphere 6.7 and the new HTML5 Client in general. But if its available all of the steps from above are orchestrated by the plugin.

Regards
Joerg

Reply
0 Kudos
aughsydney
Contributor
Contributor
Jump to solution

Hi Joerg,

Yes you are correct the NetApp was NFS and the MSA is iSCSI (block).  That's brilliant the NFS LUN is a "file" on Netapp - that makes sense.

The only thing is that its not possible to simply select the datastore and then select "Expand" in esxi (You could try, in vsphere, and use "Increase Datastore Capacity" but that would involve presenting a new MSA volume to the ESXi host - instead of what I did, increase the size of the original volume on the MSA).

I think you've answered my question!  Thanks very much.  Stars for you!

regards

p.s. You are also correct about 6.7 and the MSA2040 (not supported, I inherited this from a consultant! I might end up downgrading to esxi 6.5)

Reply
0 Kudos
aughsydney
Contributor
Contributor
Jump to solution

My last comment about the 6.7 and MSA 2040.

esxi6.7 and the MSA 2040 are not supported even with the latest release of the MSA 2040 firmware GL225-P001

Hoewever, I've no choice, somone else bought the 2040 and 6.7, I have to make it work.

I've come acorss a rather weird problem with esxi 6.7, Software iSCSI and the MSA 2040 (no idea about HW iSCSI or FC)

This is the scenario - On the MSA 2040 a virtual disk group is created (Pool A) using disks (in a disk group) that have a sector format size of 512n (see attached)

A volume is created on Pool A and presented to the esxi 6.7 hosts via iSCSI

A storage rescan is performed on esxi 6.7 - the storage device is added as a datastore (using all the default options).

At the stage in the scenario all is good - there are no problems.

However at a later date in the scenario a decision is made to add additional storage to the MSA.

Additional disks are purchased - the disks are exactly the same as the disks already in Pool A apart from one thing - the Sector Format size - the Sector Format size on the new disks is 512e.

This should not be a problem - the HP documentation says mixing and matching sector sizes is supported - hmmmm.

The scenario continues and the disks are added to a disk group in virtual disk group Pool A.

So now you have two disk groups in Pool A, each disk group is using disks of a different Sector Format size.

In this scenario an esxi 6.7 host is rebooted - upon reboot the datastore(s) located on Pool A will no longer load (You will see the device but the datastore will not mount)

After extensive testing I found that the problem was the Sector Format size.  Basically if you add a new disk group, made up of 512e disks, to a Virtual Disk Pool that already contains a disk group made up of 512n disks, and that Virtual Disk Pool already contains volumes previously presented to esxi 6.7 and mounted as datastore(s), you will find that the datastore(s) will not mount upon reboot of the esxi host.

Further testing has also revealed that esxi 6.5 is not affected by this scenario.

Reply
0 Kudos
aughsydney
Contributor
Contributor
Jump to solution

Oh - the solution.

In the MSA - when purchaseing new disks, only purchase disks that have the same Sector Format size as the disks already in the MSA.

Testing has revealed that if all the disks(SAS and SSD) have the same sector format size , adding additional disks to the virtual disk pool at a later date, will not trigger the problem describled.

Reply
0 Kudos