Sign in with
Sign up | Sign in
Your question

Setting up emergency back-up server?

Tags:
  • Networking
  • Active Directory
  • Servers
  • Network Management
  • Windows Server 2003
  • Windows Server 2008
  • Windows Update
Last response: in Business Computing
Share
July 2, 2014 12:26:22 PM

Hey all!

So I have a network here of about over 100 terminals, I have a windows server 2003 that I am running as my main file server and another server that is not being used as a hosting server like it once was. The hosting server is windows server 2008. So I want to migrate everything over to the 2008 and I can do that, but here is my question. I'm not going to throw out my old server, it can be upgraded to sever 2008 at anytime ( Other server has 2 Xeon processers, while this one only has one. Hence the wanted change) and it will probably last another 3-5 years. However what I am wondering is there some sort of way I can have this be a emergency back-up server? I can of course back up the data and the like, but what about when it comes to the directory? Is there a way to set that up, if that the other server fails for whatever reason, the old one will plop right back into it's spot?

More about : setting emergency back server

July 3, 2014 11:54:57 AM

You have licensing issues to consider here, especially when you're dealing with an extra processor. Legally, you may not be able to do what you described with your licenses.

Let's assume you can, however. Because the two differ quite a bit (one is DP, one is single), you may not be able to get the same Windows HAL to work on both. So you probably couldn't just image one off to the other, or even pull the HDs from one to the other if the motherboard/PS died. There are a few server software packages that basically replicate one server to another, even while running, but I'm not sure that would work if they weren't similar in hardware.

Second, older Xeon DP-capable processors were often slower than the singles. Older software is usually single-threaded, so having multiple but slower processors may actually be slower than a single but faster Xeon on those programs.

Finally, the major consideration is that it really isn't a good backup solution as it's old hardware as well. There is a much higher chance of something not working on older equipment than having a new, tested backup. That's why companies have hardware refresh cycles, not only for performance reasons but also to keep hardware fairly current and less likely to fail.

So I don't recommend trying to keep that old one alive as a backup. If you did, and it failed on you when you actually needed it as a backup, you bear the brunt for that failure. Sure, it could happen with a new one, but it is less likely to fail.
July 4, 2014 7:17:29 PM

I'm really going to echo 2Be on this one. While it might technically work to use the server as a backup system, it is definitely not going to be recommended for several reasons. First is compatibility. There may be quite a bit of trouble getting your software set up to work between both identically, and most likely you can't just image from one server to the other. Even pulling hard drives from one to the other, the RAID controllers may not recognize, the OS may not have drivers, or what if it is your actual array which has died or your OS that corrupted? Your backup server is of no use there.

Second, complexity. Once you figure out how to even get the backup to work, which may take a LOT of working to do, you have to now shuffle around all of your services and data between servers, reinstall stuff, license stuff, set up your networking, set up your data, configure your backups, configure failover (however you accomplish that)... long story short, you actually are more likely to experience failures or problems now because of the complexity of how everything has to be set up.

And the third that I see is hardware age and compatibility. You said you have over 100 terminals in use off these servers? What happens when things go down? Is the business basically dead in the water? If so, you HAVE to evaluate what your risk is, and what the acceptable downtime or loss of data may be. It's true that sometimes for small businesses the loss of a server for a short time is not critical. Things in the network may stop working (file sharing, hosted applications, etc.) but for most small businesses they can afford to wait a couple hours for this to be addressed. However, if EVERYTHING in your business is riding on these servers, and you have a hundred or more users suddenly down in the water, you might as well close the doors and start counting the dollars flying out the door that you are losing in productivity or lost sales, etc. Your servers are already aging systems it sounds like if they are running Server 2003/2008, so it's time to budget for new infrastructure.

Server 2012 R2 now has many built-in features for fault tolerance that older Windows just doesn't have. For example, there is already a built-in replication feature. So, if you don't need the full high-availability of a full cluster system, you can set up your primary server to duplicate to a second server. They don't HAVE to be identical, but it is recommended that they are as similar as possible (at the very least try and keep the processor the same type and preferably same generation.) Windows can also do high availability fault tolerant clustering so if one virtual machine host goes down, everything should instantly start up on your other virtual machine host. However, setting up a cluster is something for larger business and can be very expensive and complex to set up and ensure complete functionality, so replication is usually the better option for smaller businesses. It's not instantaneous failover, but you can be back up and going within a few minutes.
!