How should I configure an identical server for failover purposes?

Posted on 2014-07-28
Last Modified: 2014-08-04
OK, I'll try to keep things as simple as possible.
We have one domain.
We have a virtual host (ServerA) that has 8 virtual machines stored on it.
We have an identical host (ServerB) that has copies of all 8 virtual machines.
The virtual hosts themselves aren't accessed directly. The virtual machines are what the users access.
The 8 VMs on ServerA have static IP addresses and certain names.
The 8 VMs on ServerB which are copies of the 8 VMs from ServerA have dynamic IPs and have an underscore in front of their name.  This allows all 16 VMs to be on our network at the same time.
In the event ServerA was to fail, I want to be able to:
- remove the underscore from each VM on ServerB
- give each VM on ServerB the static IP their brother's had on ServerA
- reboot each VM and have them now be the primary, active VMs

I'm trying to minimize downtime.  I know there are solutions out there that can do all of this automatically like VMware HA but these require a SAN or vSAN which we don't have money for.
That being said, I ran into an issue when running the test below:
- shut down VM1 on ServerA
- assign _VM1 on ServerB the static IP address VM1 had on ServerA
- rename _VM1 to VM1 and reboot

When the "new" VM1 comes back up, we can't login because of a domain trust issue. This requires us to login locally, remove the VM from the domain, reboot, add it back to the domain and reboot once more.
This adds an additional XX minutes which I'd like to see if we can avoid.  
I'm assuming this is because the original VM1 was already on the domain so Active Directory doesn't like that there is another VM1 trying to login to the domain.
Can I do something to prevent this like remove VM1 from AD in the event of a failure?
Any help would be appreciated.
Question by:homerslmpson
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
  • 9
  • 9
LVL 120
ID: 40224126
Why don't you just replicate VMs  (copy, restore from Host A)

from Host A to Host B....

and if you need to turn on VMs, in the event of Host A failure, just power them up....

with no modifcations to IP Address, or Inventory name?

You will always have trust issues in AD, it's based on the machine Domain Account, not the IP Address.

Why do you want ALL VMs to be on the network at the same time, this is the issue!

Author Comment

ID: 40224159
There are network shares on most of the VMs.
I'm not sure what method of copying the VMs you are referring to but I'd think no matter the method, data wouldn't be 100% in sync.  If this is the case, a copy/restore would be missing data.
My plan as of now is to use Robocopy's monitor parameter on each VM to keep all of the shared folders in sync. While I realize this isn't syncing data real time, a value of 1 seems to keep things relatively up to date.
I suppose I could setup the hosts as follows which would reduce the restore time by 50% in the event of a failure:

ServerA:       ServerB:
VM1            _VM1
VM2            _VM2
VM3            _VM3
VM4            _VM4
_VM5           VM5
_VM6           VM6
_VM7           VM7
_VM8           VM8

Open in new window

This would increase the chance of a server failure by 100% but would reduce the time it takes to fix the issue because 4 of the 8 VMs would already be in the right spot.
LVL 120

Accepted Solution

Andrew Hancock (VMware vExpert / EE MVE^2) earned 500 total points
ID: 40224172
Purchase a copy of Veeam Backup and Replication, keep your single VMs on Host A, and replicate to Host B.

or purchase a Backup and Restore product, and restore VMs to Host B, the next morning.

What you are currently doing is prone to failure because the Windows machines are part of the Domain.

So you will have to have unique machine names and different IP addresses, and then perform "the swap" at Host down.
Industry Leaders: We Want Your Opinion!

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!


Author Comment

ID: 40224223
We already have Veeam. We use it for backing up our VMs to disk for DR purposes.
I looked at it recently as you weren't the first person to mention Veeam.  I don't recall seeing anything in that software that looked helpful.
Does replication happen real time or is it something that is scheduled daily/weekly?
LVL 120
ID: 40224256
Replication is user configurable, you create a Veeam Replication, job and VMs can replicate every 15 minutes if required to a host.

Author Comment

ID: 40225072
In the scenario I provided, am I supposed to make the entire array on ServerB a Veeam backup repository?
LVL 120
ID: 40225113
You can replicate

VMs from Host A to Host B

VMs from Host B to Host A


all VMs from Host A to Host B (standby server and VMs)

via Veeam Backup and Replication Jobs.

there is then no requirement to change machine names, IP Addresses, and in the event of Host failure, just turn on VMs on the Host.

The VMs, if set to 15 minutes, will be 15 minutes old, which is pretty good for DR.

If you require better recovery than 15 minutes you will need to look at different technology.

There is also a benefit, because once you have fixed Host A (failed host), Veeam can rollback......replicate the other way. (up to you how you work it!)

Author Comment

ID: 40225144
That sounds like a great idea but I could use a hand implementing it.
I like the idea of having all of the VMs on ServerA and replicating every 15 minutes to ServerB.
In Veeam you need to setup a backup repository.
I don't know how this should be done. Clearly we want the VMs to be stored on ServerB but I don't know how to create a repository/share using vSphere.
LVL 120
ID: 40225154
Create a Replication Job

Author Comment

ID: 40225181
I did create a replication job. It says you need to make a repository first.
I went through the steps to make one and my options were somewhat limited.
I could use local storage on the machine Veeam is installed on, a Windows network share, etc.
I don't know how to tell it to place the images on the 4TB of disk space on ServerB.
LVL 120
ID: 40225245
Local Storage, Windows Network Share is supported as a repo.

There is some confusion about the repository, this is a location where the metafiles for the backup (job info etc are stored)

later in the job, you'll get to select the datastore on the ESXI host!

where the replica files are stored (e.g. the vmdk!)

Author Comment

ID: 40226357
OK I think I understand now.
I may have been looking at BACKUP jobs not REPLICATION jobs.
I'm now running a test replication job but am getting terrible transfer speeds.
I'm replicating from one virtual host to the other and am getting a speed of 11 MB/s.
The bottleneck showing in Veeam is the TARGET.  
Any ideas why this is happening?
As of right now there is nothing on the 2nd virtual host besides the ESXi OS itself.
And thank you for your help!
LVL 120
ID: 40226423
Speed of disk, network speed.

Author Comment

ID: 40226431
They are both connected to the same 1Gbps switch.
They are both running a RAID10 w/ 8 x 1TB SAS drives.
I know they aren't SSD or anything but jeez there's nothing even on them yet.
LVL 120
ID: 40226487
We you getting better Backup speeds, this is probably worth a new question, on Veeam specific configuration.

Author Comment

ID: 40226719
Looks like this has to do with Veeam needing a proxy server on the target server.
Veeam itself acts as a proxy server on the host side but one is needed on the target side for optimal performance.
I was on the phone with Veeam for 90 minutes talking to one of their techs and after everything is all said and done, it would likely be a BAD idea to setup replication every XX minutes due to the performance hit the datastore would take.
Looks like I may need to figure something else; possibly some type of hybrid system.
LVL 120
ID: 40226750
That depends on how good the performance of your datastore is.

and how much downtime you can afford to lose....

60 minutes...

otherwise products like DoubleTake HA, are better at this....because data is replicated at the VM level, not the entire VM...

Author Closing Comment

ID: 40238667
The goal was to try and avoid purchasing additional software. The initial software purchase is one thing but you need to have active maintenance in order to get support which in some cases is quite costly and continues indefinitely.
I'm working with Veeam on the speed issue and hopefully they can help me resolve it.
Thanks for you help.

Featured Post

Edgartown IT Case Study

Learn about Edgartown's quest to ensure the safety and security of the entire town's employee and citizen data. Read the case study!

Question has a verified solution.

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

If we need to check who deleted a Virtual Machine from our vCenter. Looking this task in logs can be painful and spend lot of time, so the best way to check this is in the vCenter DB. Just connect to vCenter DB(default DB should be VCDB and using…
Giving access to ESXi shell console is always an issue for IT departments to other Teams, or Projects. We need to find a way so that teams can use ESXTOP for their POCs, or tests without giving them the access to ESXi host shell console with a root …
This tutorial will walk an individual through the process of configuring basic necessities in order to use the 2010 version of Data Protection Manager. These include storage, agents, and protection jobs. Launch Data Protection Manager from the deskt…
After creating this article (, I decided to make a video (no audio) to show you how to configure the routers and run some trace routes and pings between the 7 sites…

730 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