Solved

vSphere/VMware: File Server Cluster vs FT - need some advice

Posted on 2011-02-11
6
1,193 Views
Last Modified: 2012-05-11
Hi,

We are migrating a windows 2003 File server (on a physical server) to a Windows 2008 R2 Standard that will run on a ESX cluster.

The new server has to have the same server name and IP than the old once once the data is migrated.

We would like to implement redundancy for this server we know that we can go with FT (VMware) or build a MS Cluster.

We looking for some feedback about which approach is going could be better, basically PROs and CONs about implementing the rundancy with FT vs MS cluster.

Also could someone provide some advise or guideliness about implementing a MS cluster on a vSphere 4.1 cluster. Do we have to use RDM LUNs? I assume that the cluster nodes have to placed each in a different ESX host, etc...

Thank you in advance for your help.

 

 
0
Comment
Question by:llarava
  • 3
  • 2
6 Comments
 
LVL 37

Expert Comment

by:Neil Russell
ID: 34873463
If you have an ESXi cluster with SAN storage for your VM's and full failover enabled on the cluster then why would you want to cluster the MS side of things as well? You are talking about running a cluster on a cluster?

Just have a good solid ESXi cluster with your infrastructure on it.

Don't forget about licensing when you enable failover!
0
 

Author Comment

by:llarava
ID: 34873544
I think you got it right I want a cluster within a cluster because every type of cluster is going to provide me with a different type of redundancy.

The ESX cluster is providing physical node redundancy in case of a phyisical node failure.

FT or MS cluster will allow me to failover in case of an application failure which is complimentary to HA.
 
0
 
LVL 28

Expert Comment

by:bgoering
ID: 34876283
In a nutshell -

MSCS clustering will provide for a seperate server to pick up the load in the event the server (vm) goes down that is hosting the resource, file shares in your case. IP address for the fileshare floats amongst whichever cluster node is hosting it at the time. You would not name your new vms netbios name the same as your existing box as that name would be used for the cluster resource. See http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1004617 for more information and links to resources on implementing a MSCS cluster on vSphere. The biggest CON for this is that there is a brief outage while the share transfers from one server to another.

VMware Fault Tolerance (FT) is fairly new technology that allows a mirror vm to run in lockstep on a seperate ESX(i) host than the vm that is hosting the application. In the event the ESX(i) host goes down the mirror vm immediately picks up the load with no down time to users. Optionaly FT can be configured to start a 3rd mirror vm on yet another ESXi host in case of failure of the first. The biggest CON is that you are limited to a single processor for FT virtual machines.

Another possibility is VMware High Availability (HA). Like FT, HA protects you against a failure of the ESX(i) host. When a host failure is detected, HA will restart the vms that were running on the failing host on other ESX(i) hosts in the cluster.

Summary - FT has 0 downtime, but is limited to a single processor. Does not protect against a failure in the OS or Application running on the vm because the failure would be mirrored on the FT vm.

HA has moderate downtime as the vm must be restarted (and possibly go through error recovery) on another host. Again won't protect against an OS or App crash, but like FT only protection is for a failure of the ESX(i) host.

MSCS has generally minimal downtime. It does protect against a failure of the OS or App of a virtual machine as well as when the hosting vm disappears because of a host ESXi crash.

Personally for maximum protection with only minimal down time I would go with MSCS + HA.

Good Luck
0
Netscaler Common Configuration How To guides

If you use NetScaler you will want to see these guides. The NetScaler How To Guides show administrators how to get NetScaler up and configured by providing instructions for common scenarios and some not so common ones.

 
LVL 28

Expert Comment

by:bgoering
ID: 34876299
Just notice your statement "FT or MS cluster will allow me to failover in case of an application failure which is complimentary to HA"

Actually FT will mirror any application failure in near real time on the FT vm... as stated above. MSCS is the ONLY one of the three that will protect somewhat aganst Application and OS failure.

I don't believe you can use MSCS with FT (but not absolutely sure) because using rdm disks (as opposed to vmdk) I am pretty sure is still not supported with FT. RDM would be needed to set up MSCS.
0
 

Author Comment

by:llarava
ID: 34888006
Just notice your statement "FT or MS cluster will allow me to failover in case of an application failure which is complimentary to HA" - You are correct.

A couple of  questions:

1.) VMware FT  

FT replicates the active VM to a passive VM in order to provide redundancy in case of a node failure. We will be implementing FT on a File Server which will be created as follows:

C:\OS -> 1 single vdisk 40 GB
D:\DATA -> 1 single vdisk 400 GB

Then they both will be placed together within the same VM.

According to disk configuration if we want to enable FT does it mean that OS and DATA volume need to be replicated? In other words that we are going to have 2 VM whit a total of 440GBs?  


2.) MS cluster that will run over the VMWare ESX cluster.  

Are you aware of the requirements that are needed in order for the ESX server to support a MS Cluster? For example do we need to use RDM Luns for the QUORUM or the cluster shared drive where the data is going to be shared by the two VM cluster nodes?

0
 
LVL 28

Accepted Solution

by:
bgoering earned 500 total points
ID: 34888173
The disk is not replicated during fault tolerance - it is the running state of the machine. The mirror machine runs in lockstep with the primary machine and executes all of the same instructions. What VMware does is intercept the scsi I/O on the mirror machine and return whatever the I/O result was on the primary to the mirror, but doesn't actually do the I/O. In case of failover VMware quits intercepting the I/O and allows it to go through to the disk. But it is the same disk.

For number 2 you can download the appropriate detailed info from VMware from http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1004617 - this will point to a pdf with all of the details. The answer depends on many things, the version of windows, is it cluster in a box, across boxes, between a vm and a physical server, etc.
0

Featured Post

NAS Cloud Backup Strategies

This article explains backup scenarios when using network storage. We review the so-called “3-2-1 strategy” and summarize the methods you can use to send NAS data to the cloud

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

A procedure for exporting installed hotfix details of remote computers using powershell
When rebooting a vCenters 6.0 and try to connect using vSphere Client we get this issue "Invalid URL: The hostname could not parsed." When we get this error we need to do some changes in the vCenter advanced settings to fix the issue.
This tutorial will show how to configure a new Backup Exec 2012 server and move an existing database to that server with the use of the BEUtility. Install Backup Exec 2012 on the new server and apply all of the latest hotfixes and service packs. The…
How to install and configure Citrix XenApp 6.5 - Part 1. In this video tutorial we have explained step by step installation of Citrix XenApp 6.5 Server on Windows Server 2008 R2 is explained in this video. We have explained the difference between…

770 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question