Vsphere design for a DR and high availability scenario

[Rsant]

Hi All,

We bought four servers to create a new virtualization design thinking in DR and high availability. Today we have two sites with a distance of 500 meters between them.

The two sites are connected via fiber channel so for example I can access a LUN from the second site in a storage located on the first site via FC. The first idea is install two servers on the first site and two servers on the second site, the four servers in one cluster (HA and DRS) or one cluster (HA and DRS) with four nodes.

My issue with this configuration is if the vCenter (located on the first site) could not contact the two servers on the second site, the Vcenter will start the vmotion of the VMs to migrate the VMs to the servers located on the first site assuming that the servers located on the second site are down! It can be true or not, if the problem is a simple network problem affecting the comunication between the sites, the servers aren´t down!

I tried to explain as best as I can, but if you have any doubts let me know.

So I would like to share this scenario with you to know if you have a best solution based on previous experience or any consideration about this plan.

Thanks.

[DSTAVERT]

Where is the storage located? HA wouldn't be my idea of a DR solution. Most DR scenarios involve a human making some decision about the accuracy of the condition. If the Fibre connection is down do both locations still see the SAN? If not you have a major flaw in your design.

Forum Upgrade Notice - the VMware Communities forums will be upgraded the weekend of December 12th. The forum will be in read-only mode from Friday, December 10th 6 PM PST until Sunday, December 12th 2 AM PST.

[Rsant]

The storages are located on the first site. The servers on the second access the LUNs through FC and we have redundancy between the FC cables that connect both sites. We have redundancy on the FC switches on the tow sites too.

I know that HA is not a best solution for DR but, I believe that for now I could use this and in a near future start to study the Site Recovery Manager.

[idle-jam]

If your network is not stable (it should be with FC), then i suggest doing away with DRS or VMotion. Each site has their own cluster. Anyway does your SAN supports LUN snapshot replication? you could replicate a copy to DR and then manually mount the VM (the manual way of SRM).

iDLE-jAM | VCP 2, VCP 3 & VCP 4

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points

[DSTAVERT]

This isn't really DR just a wider local network. You need to have some form of storage in both locations and some type of replication between them. This just, for want of a better word "protects" you, if all the servers fail in one location.

Forum Upgrade Notice - the VMware Communities forums will be upgraded the weekend of December 12th. The forum will be in read-only mode from Friday, December 10th 6 PM PST until Sunday, December 12th 2 AM PST.

[Rsant]

Acctually, I have storage on both sites and the data is replicated between the storages, but my main question is:

If my VC lose the contact with the servers located inthe second site (per example), the VMs will be started through HA on the servers in the primary site?

I am saying that, because the problem can be a simple network problem and not a really problem with the ESX hosts and, is this was true, I will have the same VMs started on the ESX hosts in the primary and secondary site, and this will be a problem.

Let me know your opinion...

[AndreTheGiant]

VMware HA does not require vCenter to run.

In fact HA can also restart vCenter Server VM!

Andre

[idle-jam]

On top of of that, do check out advance ha settings where you can have second isolation address and etc. Another option for VMware vCenter resiliency would be looking into VMware vCenter Heartbeat.

iDLE-jAM | VCP 2, VCP 3 & VCP 4

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points