VMware Cloud Community
epj
Contributor
Contributor

2 Vm Farms on 2 Different Sans

Ok Lets see if I can explain this so that it makes sense even to me Smiley Happy

we currently have a Prod/dev VM datacenter with 6 ESX hosts connected to a Dell EMC CX300 with several VMFS regions carved out. Each host has dual Fibre HBA's with one leg running through one Firbre switch and the other leg running through a second switch for redundent pathing. All hosts of course can see all Luns as needed for vmotion etc. All hosts are currently at 3.01 and have most patches across the board.

Now we are building a new Prod only datacenter that will consist of at least 5 dual dual core processors, dual HBA's and connected through their own fibre switches and then to their own Dell EMC CX320 with something like 4 534GB VMFS luns and one 480GB +/- a GB or 2.

Oh also for now we have decided to stay at 3.01 and patches that are the same for the new farm as we want to move over all of the prod servers to the new farm with as little amount of issues as possible.

Of course we understand that we will have to do a cold copy over as the Hosts are different, but all other things will be configured pretty much the same where possible (I.e. Virtual switches, etc)

Both Datacenters would be controlled by the same virtual center (they just do not share SANS!)

Now to the actual question (s) at hand

1:given that the farms are isolated and on their own San box, how would be the best way to get the prod VM's moved over to the new datacenter?

We have ESX rangerPro so one thought was try to restore to the new farm do the machine convert bring online and do tools if needed, rense repeat and do it again till done! Is this even possible given the above information???

2: another thought was this: can I connect one path from each of the fibre switches on the old datacenter to the new switches and set a zone so that both farms can see each other and then do a cold migration between the 2 farms (moving the vmdk file as well to the new datacenter)?

If this was to be workable would it require a complete shutdown or interruption of the fabric ? or would this just be a simple thing like adding a new hosts and updating the zone on the fly??

Number 2 if it would work would be the best as I could if I had to restore to the old datacenter in and emergency or even move the dev boxes over to the prod farm (same reasons)

Any thoughts suggestions white papers drawings (grin) to suggest how one would get there from here would be much appreciated ....

Thanks to all you guys who put in so much time and effort on these groups!

0 Kudos
1 Reply
cmanucy
Hot Shot
Hot Shot

That shouldn't be a problem (multiple HBA's, SANs) on a single host - SO LONG AS the LUNs don't have the same name.

We're doing that now migrating from a POS FC SAN to a better iSCSI setup ... one ESX host has both the Fiber HBA and the iSCSI setup, and we do cold migrations for each of the VM's that can be down for a while.

For those that can't be down, we'll use vRanger and/or EsXpress to restore them and then deal with the data delta issue.

---- Carter Manucy