Adding addition VMFS partitions (datastores)

[p2eky]

Hi Everyone,

I am new to this forum with limited knowledge of VMWare. I am looking for sone help to save me some time!

Here is the background and configuration

Have a 4 node cluster

1. I presented a new LUN from SAN to ESX

2. Put one ESX host into maintenance mode

2. From "Configuration" tab then "Storage Adapters" i did a "Rescan then selected Scan for New Storage Adapters"

New LUN successfully showed up

3. Repeated process for all 4 servers

4. Put one ESX host into maintenance mode again

5. From "Configuration" tab then "Storage (SCSI, SAN and NFS)" i chose "Add Storage" and created a new datastore successfully.

This datastore is now visible on 1 ESX host.

My question is... For each additional ESX host do i need to put it into maintenance mode and do the following

From "Configuration" tab then "Storage Adapters" Rescan then choose Scan for new VMFS volumes or can i do it with Virtual Machines running live?

Thanks

David

[jgalexan]

You do not have to put the host into maintenance mode for this operation. After the LUN has been added to the first host, it already has a VMFS file system on it. You simply need to rescan the HBAs on the remaining hosts (I usually have to scan 2 times, don't know why), then you may need to refresh your storage view as well (configuration > storage). You should now see the additional Datastore in your list.

Hope this helps.

[VMKR9 · Accepted solution]

No, on the remaining hosts just hit refreshin the top right on the "Storage(SCSI, SAN and NFS)"

[p2eky]

Hmmm. I have roughly 10VMS on each host and would hate to kill them somehow!

Are we saying the refresh does a similar function to Rescan .... VMFS partitions action?

[Texiwill]

Hello,

None of the steps for adding a LUN to an ESX server require being placed into maintenance mode. You would not want to do that on a productions system with lots of VMs running upon it. Also be aware that there is a 1 LUN to 1 VMFS rule. Any more VMFS to a LUN and you run into horrendous locking issues.

Best regards,

Edward L. Haletky

VMware Communities User Moderator

====

Author of the book 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', Copyright 2008 Pearson Education. CIO Virtualization Blog: http://www.cio.com/blog/index/topic/168354, As well as the Virtualization Wiki at http://www.astroarch.com/wiki/index.php/Virtualization

[p2eky]

Texiwill - the reason i ask is.. a colleague added a LUN a few weeks ago and it crashed every VM on the host! Since then i have become paranoid

[jgalexan]

Was a case opened with VMware support to find the cause to the crash? Adding a new volume with a new LUN should not interupt existing running servers in yout evironment.

Rescan is done to allow the HBAs to see additional LUNs that have been presented to the host.

Refresh is done to refresh your screen.

[p2eky]

No we didnt log a call. Think people were just glad to get it back up - fortunately i wasnt here at the time!

I think the way to go is... Maintenance mode (to be on the safe side) then Rescan for VMFS partitions on each ESX host. Looks like i'll spend my evening migrating VMs around

[VMKR9]

P2eky what version of ESX are you running, there was a problem with an older release that would crash the server when you scanned all the hbas at once, cant remember the exact hardware and version that affected it but the temp fix was to only scan one hba at a time

[p2eky]

ESX v3.0.1 Patched up to November 2007

To clarify - The hosted Virtual machines crashed when the hba's were scanned. Upon restarting the virtual machines they would appear to "jump" ESX servers i.e visible on ESX01 for a few secs then visible on ESX04.

The fix was to shutdown "both" and start from fresh.

[VMKR9]

While you have the servers in maintenance mode I would patch them!

[Texiwill]

Hello,

I can understand you concern. I would also apply the appropriate patch, or if you have already or are using a newer version of ESX contact your VMware Support representative, this is NOT normal behavior and should not happen.

Best regards,

Edward L. Haletky

VMware Communities User Moderator

====

Author of the book 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', Copyright 2008 Pearson Education. CIO Virtualization Blog: http://www.cio.com/blog/index/topic/168354, As well as the Virtualization Wiki at http://www.astroarch.com/wiki/index.php/Virtualization

[BenConrad]

I've had servers hang many times while performing iSCSI rescans (one adapter at a time, not using the global 'rescan' link). For a while, we only added volumes when the host was in Maint mode. We had all the ESX QLA and iSCSi 'rescan' patches for ESX 3.0.1 at the time. Ultimately, it was a problem with iSCSI, QLA, path failover & boot-from-SAN, the COS would go read-only and we'd have to power off the ESX server. We've moved away from QLA boot-from-san and have not had any issues during rescans.

Ben