Articles

Solved Server 2003 Backup / Disaster Planning

December 13, 2012 at 21:14:04
Specs: Windows Server 2003

I currently have two identical servers, each with a 500GB SATA drive. Server1 is our production server, and if it ever goes down we are in big trouble. I would like to 'clone' the hdd on Server1 and install it on Server2. This way if Server1 goes does, Server2 goes right into production.

I did a full system backup on Server1 and tried to install in on Server2 but the system would not boot.

What is the best way to replicate server1 to server2?

Any direction offered is most appreciated.


See More: Server 2003 Backup / Disaster Planning

Report •


✔ Best Answer
December 14, 2012 at 18:13:57

Another alternative since it looks like you may be rebuilding the server using RAID, is to build SERVER2 with the new Raid1 configuration using a few drives. On this server, install VMWare ESXi and then P2V the server using Hotclone over to VMWare. I would recommend a small RAID1 setup for the VMWare installation (72GB drives or less) and then 3 500GB drives in a RAID5 config for the data store. This data store is what will store the actual VMs. Once you are up and running on the virtual server and everything is stable, rebuild the first server as another VMWare host. There is software out there to allow backing up of the Virtual Machines from within VMWare (Acronis, Veeam, etc). This will allow you to import the backed up copy of the Virtual Machine directly into the second server. once you are up and working on the backup server, you can work on migrating the data store from the main server into the backup.

The main cost for an implementation like this would be in the cost of the additional drives. VMWare ESXi is free, so there is no cost on that side. The VM backup software would also be an additional cost, but you should be using good backup software no mater what server implementation method you choose.

This method will help to ensure that even if you have a major hardware meltdown on the server end of things, that you are able to get back up and working with minimal data loss and minimal downtime.

I also agree that the servers should be connected to good quality UPS units (APC, Tripplite, Belkin). These do not have to be expensive devices, I can suggest some if you wish.

As far as network redundancy goes, with 6 people, a 16 port gigabit network switch should more than suffice. To ensure you have something to use in the event of a failure, pick up an 8 port gigabit desktop switch. These can be had for less than $120 now. It may not be the best for full time use, but if your main network switch takes a dump, it can be invaluable to have around to swap in and get things back up and working.

I also agree with Wanderer that you should have support contracts in place on both servers. Dell ProSupport would be the support plan to look into for the servers. The base ProSupport plan has next business day onsite repair times, but this can be upgraded to 8, 4, or 2 hours onsite support. 8 or 4 would be the typical support level for mission critical servers.

I have several small business clients with 5-15 users who have a similar setup to the one that I recommended above (except without the second physical server in case of a hardware failure), which is why I feel that this setup may get you the best "bang for your buck".

Please let me know if you have any questions or need more information. :)

-----
IT Desktop & Network Consultant - MOS Master Certified, MCP, MCSA, MCITP - Windows 7, CCNA Certificate Pending, A+, Network +

::geek::



#1
December 14, 2012 at 07:35:50

when you say "each with a 500 GB SATA drive" do you mean each of the two servers only has a single HDD in it?

If this is the case, you need to start by building redundancy IN the servers. Which is to say, buy a good hardware RAID controller for each server and some more drives and build some RAID's in your servers. My preferance has always been a RAID 1 for the OS and a RAID 5 for the data. There are other good combinations like a RAID 10 for both OS and data and hot spares etc.

But the idea of cloning the entire box is moot if you have the RAID redundancy within the server. Then you could use the second server (also RAID'd) as a secondary DC to provide redundancy for your domain.

It matters not how straight the gate,
How charged with punishments the scroll,
I am the master of my fate;
I am the captain of my soul.

***William Henley***


Report •

#2
December 14, 2012 at 07:57:29

You should be able to clone the drive between the systems. What cloning software are you using? It has to be a server product not a workstation product to work.

Also realize that given the scenerio of one server going down and you bringing up the other server that you have lost data, EVEN if you restore server2 from the most recent backup of server1.

This is where you need to consider clustering and using a backend san. Then you have system failover with no data loss.

But your first step is to get those drives raided asap. No one that cares about their data runs on a single drive. I don't, even with my home systems.

Answers are only as good as the information you provide.
How to properly post a question:
Sorry no tech support via PM's


Report •

#3
December 14, 2012 at 09:33:36

Thanks to you both for your reply. To answer a couple of your questions: Yes, each server only has on SATA drive each; with the production server having a tape backup doing nightly incremental backups. I have not tried to clone it as of yet, but in the past I have used Norton Ghost, but admittedly I have never done it with a server.

I do have a RAID controller for one of the servers (the servers are the older Dell PowerEdge 750), but I took it out and installed the SATA instead. I somewhat understand what you are saying about building Raids on the server, then what happened in the even of a major hardware failure such as a motherboard? I would then just install the drives on Server2 (hoping that the OS and Data is still in tact)? Also, in order to accomplish this I would have to completely rebuild the server, correct? (install new os and then add applications/data and then configure everything the same as the Production server?)

Just as background - Data is not the most important thing about our server, rather it is the applications that we run. We have EDI exchanges with other companies and this function is far more important than the data (given that the data is on the nightly tape backups for data).

Just so I understand, you still feel that Raid contollers are the way to go?

Thanks again


Report •

Related Solutions

#4
December 14, 2012 at 09:50:47

"nightly incremental backups"

Bad choice for backups. I only recommend full nightly backups which include system state.

If operations are more important then your server design is completely wrong for the task.

Servers should be in a cluster configuration
Servers should have 4x7x365 maintenance contracts with 4 hr response time
Servers should have multiple power supplies and be in a raid1 configuation minimum [raid 10 better]
Servers should be on UPS's with a serial connection/software monitoring
Backbone switch should have redundancy builtin [multiple power supplies/ switched fabric modules so if one "brain" goes out the other takes over.
Dual wan port router with cable for one and dsl for the other so you have failover and are using dissimilar providers.

Most importantly you need a documented disaster recovery plan.

Your present setup is for a mom and pop shop setup not a critical business setup.

Answers are only as good as the information you provide.
How to properly post a question:
Sorry no tech support via PM's


Report •

#5
December 14, 2012 at 10:36:19

Thanks again for your reply ... you have exposed us for what we truly are: a "mom and pop shop" with aspirations of having stable data solutions. We are only a 6 person office and being the owner I am also the main IT guy. I have helped setup servers at my former occupation and understand what you are saying about the “critical business setup” but the reality is that we are not a multi-national with a huge budget.

It is obviously though that we have a lot of exposure to risk here so will I will work to rebuild using Raids with OS and data separately. I was hoping just to clone the two servers so that if one goes down I could just swap it out with the other - that sounds like it will not work for us. I can afford to go down for an hour … I cannot afford to go down for a couple of days.

Thanks again for your comments


Report •

#6
December 14, 2012 at 14:58:01

Cloning the server is doable.

You just need server cloning software since most key their software to either workstation [sometimes free] or server [never free and can be expensive.

You do want to do the raid1 for both and full backups which include system state. That and a good firewall to protect from the bad guys and you should be set.

Nothing wrong with a Mom and Pop shop. I have had a few of my own though those were Pop only :-)

Answers are only as good as the information you provide.
How to properly post a question:
Sorry no tech support via PM's


Report •

#7
December 14, 2012 at 18:13:57
✔ Best Answer

Another alternative since it looks like you may be rebuilding the server using RAID, is to build SERVER2 with the new Raid1 configuration using a few drives. On this server, install VMWare ESXi and then P2V the server using Hotclone over to VMWare. I would recommend a small RAID1 setup for the VMWare installation (72GB drives or less) and then 3 500GB drives in a RAID5 config for the data store. This data store is what will store the actual VMs. Once you are up and running on the virtual server and everything is stable, rebuild the first server as another VMWare host. There is software out there to allow backing up of the Virtual Machines from within VMWare (Acronis, Veeam, etc). This will allow you to import the backed up copy of the Virtual Machine directly into the second server. once you are up and working on the backup server, you can work on migrating the data store from the main server into the backup.

The main cost for an implementation like this would be in the cost of the additional drives. VMWare ESXi is free, so there is no cost on that side. The VM backup software would also be an additional cost, but you should be using good backup software no mater what server implementation method you choose.

This method will help to ensure that even if you have a major hardware meltdown on the server end of things, that you are able to get back up and working with minimal data loss and minimal downtime.

I also agree that the servers should be connected to good quality UPS units (APC, Tripplite, Belkin). These do not have to be expensive devices, I can suggest some if you wish.

As far as network redundancy goes, with 6 people, a 16 port gigabit network switch should more than suffice. To ensure you have something to use in the event of a failure, pick up an 8 port gigabit desktop switch. These can be had for less than $120 now. It may not be the best for full time use, but if your main network switch takes a dump, it can be invaluable to have around to swap in and get things back up and working.

I also agree with Wanderer that you should have support contracts in place on both servers. Dell ProSupport would be the support plan to look into for the servers. The base ProSupport plan has next business day onsite repair times, but this can be upgraded to 8, 4, or 2 hours onsite support. 8 or 4 would be the typical support level for mission critical servers.

I have several small business clients with 5-15 users who have a similar setup to the one that I recommended above (except without the second physical server in case of a hardware failure), which is why I feel that this setup may get you the best "bang for your buck".

Please let me know if you have any questions or need more information. :)

-----
IT Desktop & Network Consultant - MOS Master Certified, MCP, MCSA, MCITP - Windows 7, CCNA Certificate Pending, A+, Network +

::geek::


Report •

#8
December 14, 2012 at 18:44:13

Song Cloud, thank you very much for your informative response. All three of you have convinced me to rebuild a server using Raid (and VMware). I will also look into support contracts as well.

Thank so much to all of you for extending me some of your expertise in this area.


Report •

#9
December 15, 2012 at 13:11:00

Another option for backing up VM's on ESXi is the "Ghetto Backup Script" if your budget is tight. It's a simple shell script that can be scheduled with Cron. There is also a free version of Veeam but I do not believe it supports scheduled backups.

Tony


Report •

#10
December 15, 2012 at 20:27:50

Hello again. For an interim solution I used Clonezilla to clone the SATA drive in Server1 over to the additional SATA drive that I had on hand. I installed the cloned SATA drive in Server2 and it works like a charm … so I can now at least sleep at night knowing that I have a backup should I incur a colossal failure in Server1. However I come to realize through your comments above that this is really quite amateuristic and not appropriate for “a critical business setup” (as mentioned by wander)

I have ordered two new SCSI Drives for my Raid Controller and will rebuild Server2 based on Raid1, as our data requirements are very minimal but our uptime is absolutely critical. As our business grows and our data demands increase we will move to a RAID1/RAID5 combination (with a new server).

Thanks to everyone who commented – all of your responses have been very insightful and opened my eyes to ensuring we have a reliable disaster plan.


Report •

#11
December 15, 2012 at 22:12:21

Glad we could help!

-----
IT Desktop & Network Consultant - MOS Master Certified, MCP, MCSA, MCITP - Windows 7, CCNA Certificate Pending, A+, Network +

::geek::


Report •

#12
December 16, 2012 at 08:20:24

Before jumping over to VMware, you need to first ascertain whether or not it will run on your hardware. Even if it does, you also have to remember there's a learning curve there too.

gateway1

The biggest concern with cloning is the hardware. If it's not exactly identical, you will have issues.

If it IS exactly identical hardware then a clone is doable.

Considering all you've said, if they're identical and you can do a hardware RAID on one of them (software RAID is not worth the effort) and you can't run VMware on your servers then I would consider running the RAID on your production server and keep the non-RAID server as your backup.

In the case of a catastrophic hardware failure, you would then have the option of swapping the RAID controller and drives over to the backup server, or swapping the bad piece of hardware for a good one from the backup server.

In either case, downtime is less than one hour and depending on what piece of hardware fails, could be less than 15 min's too.

If you have some extra $$$, you might want to give some thought to buying a NAS device as well and have your backups go to it as well as the tape. Tapes can be taken offsite, the NAS can remain in the building.

As for a UPS, I highly recommend APC. APC offers a lifetime warranty and something in the area of a SmartUPS 1500va would do what you need nicely.

Rather than spend money on expensive backup software which you'll again have to learn how to use. I could provide you a link to a guide on using windows native ntbackup (it's included wtih the operating system) to perform the backups you need to do. It's rather easy to learn and if you can't use VMware (or opt not to as the case may be) will easily do what you need it to.

It matters not how straight the gate,
How charged with punishments the scroll,
I am the master of my fate;
I am the captain of my soul.

***William Henley***


Report •

#13
August 20, 2013 at 23:52:59

Maybe you can try to use some free backup software which can provide you system backup, partition and disk backup.
http://www.aomeitech.com/

Report •


Ask Question