Steve_Douglas
Contributor
Contributor

VCB Proxy - When attached to VMFS, Server fails on reboot !

Hello All,

We have a strange issue whilst setting up VCB on our VCB Proxy server which results in the server failing on reboot... ie, the Windows 2003 screen comes up, then instead of getting to the blue startup screen "Loading Settings"..etc, the server screen goes blank and hangs. Only hard reset will get me out of this.

To setup VCB I have installed VCB on our Windows 2003 with SP" server. This server is also our Backup Exec 12.5 server, and once VCB is setup OK I will install the Integration Module, and perhaps the AVVI agent. So, anyway, VCB was installed and I also performed the "diskpart" cmmands for automount disable and scrub. I then presented the SAN logical drives for the VMFS shared volume to the VCB Proxy server as Read Only. I rebooted the server and all was well. I then tried some tests with VCB using batch command s for VCBMounter.exe and all worked well appart from som certificate errors, but they did not stop me doing "fullvm" image backups of the virtual machines. This was great, I thought I had cracked it, until I rebooted the server again and the server didn't come back up. We had initially though we had had a drive failure, then and HBA failure. We removed the HBA attached to the SAN VMFS volume and the server came back up. We had a spare and replaced the HBA. Then configure Logical drive presentation again to the new HBA's WWN, including the VMFS volume. All was well, we were back on track.... rebooted the server and it failed to start. I unplugged the fibre on the connection to the VMFS volume and hard reset server and it came back up. I removed the logical drive presentation to the VMFS volume, plugged back in the fibre and rebooted a few times and everything was fine again, whereby I could see other NTFS volumes that I had presented from the same SAN and another SAN.

It seems every time I present the VMFS volume to my VCB Proxy server I can reboot the server once, but another time and the server hangs at startup and requires a hard reset.

Does anyone know if I need to give VCB Proxy Read Write access to the VMFS volume? The documentation only states to give access, and I figured Read Only would be enough.

Has anyone got any ideas what might be happening? I read some article about the multiple paths in the SAN setup can cause problems, but now the same problems we are experiencing.

Some Tech Info:

ESX Vi 3.5 on 2 physical HOSTS

Virtual Centre 2.5

MSA 2012f is where the VMFS volume is shared (2TB)

VCB Proxy:

HP DL380 G3

2 GB RAM

Win 2003 SP2

HBA 1 - QLA2340

HBA 2 - QLA2340

0 Kudos
11 Replies
AndreTheGiant
Immortal
Immortal

Have you disable the automount function in Windows using diskpart?

Andre

**if you found this or any other answer useful please consider allocating points for helpful or correct answers

Andre | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
Steve_Douglas
Contributor
Contributor

Yes, I have used diskpart to disable the automount and I scrubbed the history also, before i presented the VMFS volumes.

0 Kudos
AndreTheGiant
Immortal
Immortal

I've got a similar problem with a physical machine connected to an FC SAN.

The only solution was to upgrade the HBA firmware and upgrade the windows drivers.

Are you using vendor drivers or Windows drivers?

Andre

**if you found this or any other answer useful please consider allocating points for helpful or correct answers

Andre | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
Steve_Douglas
Contributor
Contributor

Hello Andre,

I am using the following HBA drivers for the QLogic QLA2340.

Provider: QLogic

Driver Date: 15/02/2008

Driver Version: 9.1.7.16

Digital Signer: Microsoft Windows Hardware Compatibility Publisher

Think these are pretty up to date as I was looking into this not that long ago for attaching the server to the SAN in the other building.

Will check out if there are any updates tomorrow.

Thanks for the advice so far. What was the setup when you were getting the errors? Is there a post I can read??

cheers,

0 Kudos
AndreTheGiant
Immortal
Immortal

>What was the setup when you were getting the errors? Is there a post I can read??

There isn't a post. It was a problem in a physical environment (but with SAN connectivity).

Andre

**if you found this or any other answer useful please consider allocating points for helpful or correct answers

Andre | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
Steve_Douglas
Contributor
Contributor

OK thanks.

I have now updated the driver to version 9.1.7.18 from 9.1.7.16.

Going to try it out again today.

cheers,

0 Kudos
Steve_Douglas
Contributor
Contributor

Hello All,

After updating the HBA driver to the latest version I then presented the LUN of the VMFS volume to my VCB Proxy server. I then rebooted, and the server failed at startup with the screen going blank as before.

I left the server in this state and removed the Read Only presentation of the VMFS LUN. To my surprise, the server then continued to start up without any intervention from me.

I am going to investigate the BIOS settings for the HBA and see if there are any settings in here that are relevant.

Can anyone clarify whether the VMFS LUN should be Read Only to the VCB Server or Read Write to the VCB Server???? I am nervous of making this Read Write as I would be in bother if I lost the current VMFS volume. Smiley Sad

0 Kudos
JoJoGabor
Expert
Expert

That is strange_. Can you do some tests with a test LUN? Just create a small LUN, present it to VMware, format it with VMFS and only present this LUN to the VCB proxy as read/write - see what happens.

What SAN are you using? Is it on the HCL?

Steve_Douglas
Contributor
Contributor

Hello,

It is an HP MSA 2012fc SAN. I believe it is on the HCL and was specified by an external company as being on the HCL before installation. Just checked, and it is there.

The HBA (QLA2340) is also on the HCL.

My next step was to create a new V_Disk on the same SAN and present it... or Map it, as HP call it on this SAN, to the ESX Servers. Format it as VMFS v3 and then present it to the VCB Proxy as Read Write and see if i get the same issues at boot.???

Keep you posted.

0 Kudos
Steve_Douglas
Contributor
Contributor

Found this..... going to try it tomorrow.... yikes!!

http://communities.vmware.com/thread/168593

:smileyshocked:

0 Kudos
Steve_Douglas
Contributor
Contributor

Here is what I did....

I created a new LUN and presented it to my 2 ESX servers as shared RW storage and formatted as VMFS v3.

I created a VM on this new VMFS volume.

I then presented this to my VCB Proxy server as RW (baring in mind I have already performed the Diskpart automout disable and scrub). In Disk management I rescaned disks and the volume appeared us healthy unknown partition. I then rebooted the server.

Server came up OK. So I rebooted again. It came up OK again.

I performed a VCB fullvm backup of my test VM, and this went OK.

I then set the VMFS volume presentation to the VCB server to Read Only and restarted the server. It came up OK.

I performed the same VCB backup of the test VM, and it completed successfully.

I created another Test VM on the new VMFS volume. I have tested the first test VM and all seems OK.

I am now going to apply this to my Production VMFS volume as I am satisfied that I only need RW access for the VMFS volume for first reboot of the "Windows" VCB server. After first reboot I am going to change this back to RO for safety sakes.

The documentation on this part of VCB setup says...... "

The VCB proxy has a read‐write access to the SAN LUNs that contain either VMFS volumes (datastores) with virtual disks, or virtual compatibility RDMs you plan to back up. Do this by adding the VCB proxy to the same fabric zones where the ESX Server system belongs.

Didn't see this before... :smileyangry:

Anyway, it seems to work just as well with RO after first boot..... so far.

0 Kudos