Link to home
Start Free TrialLog in
Avatar of MichaelBalack
MichaelBalackFlag for Singapore

asked on

How to migrate the physical MS Clustering to VMware virtualization?

This is in planning to setup a new VMware-base virtualization that comprises of 4 ESXi 6.5 hosts. All of the current physical servers, some SuSE and remainder MS Windows servers have to be migrated to virtualization. There are few pair of dual-host cluster using MS Windows clustering with shared storage. I heard that you actually can forget about migrate the clustering functionality, indeed, use the VMware's features, such as (but not limited to), HA, vMotion, or even vsan, and other; to "take over" the ms clustering.  Is that true? What would be the best choices to migrate this dual-host MS Windows Clustering into VMware virtualization and how?

thanks in advance
Avatar of kevinhsieh
kevinhsieh
Flag of United States of America image

Why do the existing clusters exist? What services are they supporting?

Are your 4 new VMware hosts in a cluster? If they are, then they provide HA for VMs, such that if there is a problem with a host, the VMs will crash and get rebooted on another host. Downtime would typically be a few minutes, especially if the systems can automatically recover after the crash without manual intervention.

If a few minutes of downtime is okay, then you probably don't need Microsoft OS based clustering.

If you can't afford the few minutes of downtime, or need to be able to do maintenance of a cluster node while keeping the service up, then you still need to keep Microsoft clustering.

There is VMware Fault Tolerant VMs, which runs two identical copies of a VM on separate hosts. It has essentially zero downtime to failover, but has strict hardware, storage, and networking requirements, and any VM maintenance such as OS patching will still take down the service.

VSAN is just a way to get shared storage so you can have a VMware cluster instead of using external NAS or SAN.
Avatar of MichaelBalack

ASKER

Hi Kevinhsieh,

Thanks for the prompt sharing.

This one looks the one,

        There is VMware Fault Tolerant VMs, which runs two identical copies of a VM on separate hosts. It has essentially zero downtime to failover,   but has strict hardware, storage, and networking requirements, and any VM maintenance such as OS patching will still take down the service.

Does this make sure of the feature called "Fault Tolerance"? if so, how to implement?
I am sorry, but what is the question? You should determine what the needs are, and then what solution best meets the needs. Finally you determine what is required to implement the solution, and make sure you can get the resources to implement it.

If your cluster is for SQL. and you need 24 x 7 availability including during patching, then FT is not the solution you need, as it doesn't allow you to do maintenance. If you can get maintenance windows, but need to avoid unplanned downtime, then FT may be a good solution, though native SQL HA technologies are probably better.
This is in planning to setup a new VMware-base virtualization that comprises of 4 ESXi 6.5 hosts. All of the current physical servers, some SuSE and remainder MS Windows servers have to be migrated to virtualization. There are few pair of dual-host cluster using MS Windows clustering with shared storage. I heard that you actually can forget about migrate the clustering functionality, indeed, use the VMware's features, such as (but not limited to), HA, vMotion, or even vsan, and other; to "take over" the ms clustering.  Is that true?

Not really. VMware has functions that could be used to substitute MS Clustering.....

e.g. if you enable VMware HA, and have a single VM, the VM would be restarted within on another host should that host fail.

But HA generally would cause 1-2 minutes of downtime ? <---- is this good enough for you ?

VMware FT - remember if there is a fault in the OS, this also gets replicated to the secondary VM.

Basically this is Primary VM in sync with Secondary VM. (on another host).

Should the Host Fail, the Secondary VM will "kick in" - zero downtime.

Most still implement MS Clustering using virtual nodes (machines).
Hi Kevinhsieh,

This is not a cluster for sql. It is a cluster that host some server services, with some cifs mounting on shared storage (SAN). Clients can access for the services via vip (virtual IP address). Multiple "instances" are running in node a, and some others in node b. So, this is a active-active cluster. If a instance is down original on node 1, then, this instance will failover to node b, within 10 seconds.

FT should be the solution for this cluster. BTW, where can I get the documentation on this FT?
Hi Andrew,

No, HA is not suitable as the affected VM has to be restarted.

VMware FT should be the one to explore.
ASKER CERTIFIED SOLUTION
Avatar of Andrew Hancock (VMware vExpert PRO / EE Fellow/British Beekeeper)
Andrew Hancock (VMware vExpert PRO / EE Fellow/British Beekeeper)
Flag of United Kingdom of Great Britain and Northern Ireland image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of robocat
robocat

>I heard that you actually can forget about migrate the clustering functionality

I'm not sure if this is what you're actually asking/thinking, but you can do MS clustering on VMWare. It just needs to be configured carefully so that e.g. the cluster nodes never end up on the same physical server.

From what you describe in your use case, this might be the way to go. VMware offers lots of HA features but they don't replace a cluster.
Thanks for both experts - Andrew and robocat. At the end, we forgot about having the clustering on vm, and having HA to move the affected vm to another host.

thanks,