Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

We are in the process of a network redesign and I am now down to changing the IP address of the Equallogic iSCSI SAN that is in use by our VMware ESX servers.

The Equallogic has three active ethernet ports servicing iSCSI requests.

I'm lookingfor a solution that will allow me to renumber the interfaces on the iSCSI SAN, and reconfigure the ESX servers to point to the new addresses, without any downtime.

has anyone preformed this sort of task before? Is what I am asking even possible?

share|improve this question
add comment

2 Answers 2

up vote 1 down vote accepted

I doubt that this will be possible with Equallogic, at least not without some downtime. The Equallogic architecture will allow you to change the ip-addresses of the interfaces without any downtime (provided the new addresses are visible to the old addresses) but the Group IP-address that is used for management and as the iSCSI target address cannot be changed without some downtime - once you change that you will lose connectivity to the entire set of arrays until you change the target settings on all of your ESX hosts and rescan. I would call Dell and ask them about this but I'd be surprised if it can be done unless they've added something new recently. If you do find out that it can be done, please let us know.

share|improve this answer
2  
After preforming the migration last week I can confirm that this can not be done without some down time -- kind of :-) You are correct that once the group IP address is changed then all ESX server will lose connectivity to the SAN. However, within each VM I have set the disk timeout to 120 seconds. This then gave me a two minute window to point the ESX servers to the new group IP address, and rescan. After the rescan connections were reestablished to the to new IP address and everything was OK. I was able to to 3 ESX servers in the 2 min. window, but it was tight. No VM crashed during this :-) –  Richard West Dec 7 '09 at 17:20
    
That's an excellent piece of Sys Admin work under pressure. –  Helvick Dec 7 '09 at 19:10
add comment

Approximate steps would be:

  • Configure LUNs to be available on multiple NICs with multiple IPs on the SAN
  • Configure ESX hosts to see those LUNs via both IPs (MPIO)
  • Verify that the ESX hosts can flip across to the second SAN route with a test LUN + VM
  • Flip the traffic across for the production VMs, reconfigure the NIC that is now unused
  • Repeat for the other NICs as required

This process depends on your SAN, ESX and Switch capabilities being able to support a full MPIO configuration, and being configured appropriately.

Depending on your SAN model, this guide may be appropriate: http://www.federalappliance.com/fedapp%5Fvmware%5F3.0.1%5Fequallogic.pdf http://pubs.vmware.com/vsp40%5Fi/wwhelp/wwhimpl/common/html/wwhelp.htm#href=iscsi%5Fsan%5Fconfig/c%5Fequallogic%5Fstorage%5Fsystems.html&single=true

After a bit more reading, some say ESX 3.5's software iSCSI initiator doesn't support MPIO, while ESX 4 does. We have MPIO implemented on ESX 3.5 over iSCSI with the Software initiator, for a Clariion unit... so YMMV.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.