Start a Conversation

Unsolved

This post is more than 5 years old

6234

February 8th, 2012 11:00

Hyper-V cluster on VNXe3100 iSCSI

Hi all,

Could you help me out, situation:

Two servers connected via iSCSI to a VNXe3100. on the servers Hyper-V Windows 2008 R2 and cluster, with quorum disk on VNX.

MPIO is enabled/installen and still the VM will crach after a Storage Processor reboot.

All is connected according the VNXe configuration guide for Hyper-V cluster.

Has anyone run into this problem before?

Cheers!

2 Intern

 • 

126 Posts

February 8th, 2012 11:00

Mike,

Unfortunitly I have not done a Install with VNXe and Microsoft Hyper-V

but I did find this document,Maybe this will help you

EMC VNXe series

Using a VNXe system with Microsoft Windows Hyper-V

http://powerlink.emc.com/km/live1/en_US/Offering_Technical/Technical_Documentation/300-010-552.pdf?mtcs=ZXZlbnRUeXBlPUttQ2xpY2tDb250ZW50RXZlbnQsZG9jdW1lbnRJZD0wOTAxNDA2NjgwNWQ1MzFmLG5hdmVOb2RlPVNvZndhcmVEb3dubG9hZHMtMg__

February 8th, 2012 12:00

Hi Mike,

I have heard similar situation with VNXe and Hyper-V. I have contacted with our colleagues to see how solved the problem.

Regards

Francisco

February 8th, 2012 13:00

Hi Mike

I have found this Primus:

How to configure Windows multipathing on a VNXe 

VNXe3100/3300 failover does not work correctly in multipath environments unless specific settings are configured. If the Windows settings are not configured correctly, VNXe storage resources only permit Read-Only access if a system storage processor (SP) goes down.

Please check the EMC solution emc269809 to follow the steps and solve the issue.

Regards.

Francisco

86 Posts

February 8th, 2012 17:00

I have emailed the author of this primus as there are a few steps missing in the setup of the MS iSCSI initiator.

Specifically, he was missing out on configuring the correct source initiator for each subnet which needs to be set in the MS iSCSI initiator Advanced Tab. See page 148 for W2K3 and W2K8 and page 154 for Windows 2008 R2 of the "EMC Host Connectivity Guide for Windows" available on http://powerlink.emc.com  for detailed instructions.

11 Posts

February 9th, 2012 01:00

Hi Mich, William and Francisco, thanks for all your help, i think we have this one solved!

Great job!

2 Posts

February 19th, 2012 19:00

Hi Mike,  I am having the same issue.  What was the solution?  I am unable to find EMC269809 anywhere in the powerlink website.  Please advise.  Thanks Stephen

11 Posts

February 21st, 2012 09:00

Hi Forsakensa,

The solution for this customer next to all the imput above,

We Created two iSCSI servers per SP on the VNX, so we moved away from 1 iSCSI server with 2 IP adresses per SP.

The Quorum disk was totaly reformated and with this we recreated the cluster information.

This did the trick ;-)

Cheers

2 Posts

February 21st, 2012 09:00

Hi Mike

When you say you created two iSCSI servers per SP. How do you associate the storage with all the iSCSi servers. From my side when creating a lun you can only pick I storage server for that lun, so how do you associate with all 4 . I am a little confused here, even though the concept you mentioned sounds correct.

Regards,

Stephen

4 Posts

May 17th, 2012 07:00

i am looking for solusion to this problem :

"Known problem on VNXe connecting to Windows Cluster servers"
spacer
spacerspacerspacer

ID: emc289415
Usage: 1
Date Created: 03/09/2012
Last Modified: 04/13/2012
STATUS: Approved
Audience: Support

Knowledgebase Solution

Question: Known problem on VNXe connecting to  Windows Cluster servers
Environment: Product: VNXe Series
Environment: Product: VNXe3100
Environment: Product: VNXe3300
Environment: OS: Windows Server 2008 R2
Environment: EMC SW: VNXe  Operating Environment (OE)  2.0.x
Environment: EMC SW: VNXe  Operating Environment (OE)  2.1.x
Problem: When a storage processor (SP) is  rebooted, cluster disk goes offline.
Problem: During upgrade the virtual machines (VMs)  on the HyperV may display a stop error on a blue screen.
Problem: When there is a SP reboot/failure, IO is  not taking the alternate path even though MPIO or MCS is  configured following EMC documentation.
Problem: Cluster Disks go offline (SQL Cluster or  Hyper-V cluster) when SP is rebooted.
Fix: With the following configuration it has  been determined that a data unavailable (DU) condition may result when a storage  processor (SP) fails over: 
  • iSCSI server on VNXe with iSCSI storage (Generic iSCSi or Hyper-V).
  • iSCSI servers with two interfaces (IP addresses).
  • Multi node cluster on Windows Server 2008 R2 with HyperV.
  • MPIO or MCS configured on all cluster nodes.
  • Cluster Shared Volume (CSV) on the cluster.

The VNXe failover period seems to be long enough to disconnect disks on  hosts. During tests in the lab it has been identified to be a problem in the  cluster environment rather than in standalone setups.  Currently (March 12,  2012) there is no resolution or workaround available to resolve this problem and  there is no estimated time frame for a fix for this issue to be made available.  Product Management has been informed about this and work is still in  progress.

Currently (April 13, 2012), this is a problem with VNXe handling of  Persistent Reservations in Windows Cluster environment.  This affects all  applications set up with Microsoft Cluster Services (MSCS), for example -  Hyper-V, SQL, File Sharing,and the like.

This does not disrrupt the working of Windows Cluster with VNXe. Many  customers have been using the VNXe with this setup. The only caution is  to not reboot the storage processors on a VNXe (as part of troubleshooting or  for upgrade activity). The failover of services will cause iSCSI storage to  disconnect.

Notes: This problem has been reported in Windows  environments only, not in VMware ESX or a Linux/UNIX configuration of  MPIO.

spacer

4 Posts

May 17th, 2012 07:00

does anyone know how to get EMC269809 i can't find it anywhere .

2 Intern

 • 

126 Posts

May 17th, 2012 07:00

Is that a primus solution you are looking for..

Giuve me your E-mail adress and I will send it to you

2 Intern

 • 

126 Posts

May 17th, 2012 07:00

Avishay,

I would Call CS and see if you can contact the autor to the Primus post.

That would be the best place to start

138 Posts

August 26th, 2013 13:00

The above mentioned problem (bug) has been permanently fixed in OE version MR4 (2.4.0.20932) and later.

1 Rookie

 • 

45 Posts

September 2nd, 2015 06:00

Did you ever get a response to your question. I am also wondering the same thing.

No Events found!

Top