Distaster Recover - EVA - Continuous Access

[-stijn-]

I do have 7 vmfs3 LUNs with a size of 750GB (containing 10 to 15 VM each) that I recplicate synchronous with HP CA (Continuous Access) between 2 EVA6000 (SAN) systems.

In the near future I have to failover the LUNs to the non active EVA system. If I understand it well, a simply rescan of the storage will not work (because the paths are changing). But what else is changing?

Storage SAN vmfs3 paths? I guess only the paths will change, but all the rest will be transparent? (VIC will still know where to find a VM when you power it on?)

We do have a lot of VM using "Mapped Raw LUNs". So these paths will change as well? But is this transparent when you boot up a VM or do I have to manually change those paths in the VM config settings? I guess that this will be quite difficult if we do have a lot of raw device mappings of the same size?

We are using DRS, I guess that a VMotion will not happen to an ESX server that hasn't rebooted yet (because it doesn't now the new path yet)? Or do I have to dissable DRS?

Or there any other important things that I have to know? Are there any important settings on a VM or ESX server needed?

I read something about the following settings:

LVM.EnableResignature = 0 (default Setting)

LVM.DisallowSnapshotLUN = 0 (default Setting is 1)

Thanks

[-stijn-]

Seems that nobody is using EVA and/or Continuous Access ??

Anybody else that has experience with EMC and SRDF (same principle), or something else?

[bertdb]

a rescan will find the LUNs, you'll have to add the VMs residing on the LUNs to the inventory (vmware-cmd from the CLI or browse datastore/select .vmx/add to inventory via the VI Client). After that, your ESXes on the DR site can power up the VMs.

[VirtualKenneth]

Check my thread for full details and steps about CA.

http://www.vmware.com/community/thread.jspa?threadID=66862&tstart=0

Fyi…if you find this post helpful, please award points using the Helpful/Correct buttons.

[frankdenneman]

In the near future I have to failover the LUNs to the

non active EVA system. If I understand it well, a

simply rescan of the storage will not work (because

the paths are changing). But what else is changing?

It will work, all you have to do is to present the replicated luns with the exact lun id as the source lun's to the ESX hosts.

Storage SAN vmfs3 paths? I guess only the paths

will change, but all the rest will be transparent?

(VIC will still know where to find a VM when you

power it on?)

The paths will change, but it will be a transparant change. As the LUN ID, UUID and datastore name match the original ones.

We do have a lot of VM using "Mapped Raw LUNs".

So these paths will change as well? But is this

transparent when you boot up a VM or do I have to

manually change those paths in the VM config

settings? I guess that this will be quite difficult

if we do have a lot of raw device mappings of the

same size?

Sorry can't help you with that one. I can test this for you

We are using DRS, I guess that a VMotion will not

happen to an ESX server that hasn't rebooted yet

(because it doesn't now the new path yet)? Or do I

have to dissable DRS?

After a rescan, all disks will be available again and DRS will work.

Or there any other important things that I have to

know? Are there any important settings on a VM or ESX

server needed?

No just present the disk with the same LUN number. You don't even have to change the host access type on your eva.

I read something about the following settings:

LVM.EnableResignature = 0 (default Setting)

LVM.DisallowSnapshotLUN = 0 (default Setting is 1)

Thanks

You do not need to change those settings. The vdisks are being replicated between two controllers of the same sort, the UUID will remain intact. ESX won't see the difference.

I have a customer where we are replicating all disks between two eva 8000's. Due to a physicall move of the eva 8000's we have tested the failover and it worked perfectly.

P'm me for more details.

P.S. don't give the points to Kenneth he already has enough points.

[VirtualKenneth]

P.S. don't give the points to Kenneth he already has

enough points.

Well Frank, I do hope you get some points for story. I'm not that hateful 😛

[frankdenneman]

Thats why I love you so much babe! :smileygrin:

[alhamad]

Hi, I have similiar post about EVA replications between DR sites, it contains a link to an HP whitepaper that explains the process. Please check it out:

http://www.vmware.com/community/thread.jspa?messageID=716720&#716720