Jump to content

Recommended Posts

Posted

Just creating a disaster recovery plan of actions list, should the worst happen.

 

This primary school have 2 servers (2008r2). The main server (DC1) has all the roles, all the services and all the data. The second sever (DC2) is just a Global catalog DC with nothing else.

 

The main server also has a good backup with me doing full bare metal backups onsite and in addition system state + data going remotely off site.

 

So should the worst happen and the main server goes down and needs to be replaced / rebuilt, what would be my best plan of action?

 

Would it be something like, move the roles to DC2, install OS on new sever using same IP and name as old server, migrate roles back from DC2, install all software and then restore data (staff / pupils work) from backup.

 

or would it be something like, do a bare metal restore.

 

Not being able to actually test this out completely I would very much welcome any recommendations or issues I should be aware of.

Posted
What happens if both of your servers (which i presume are in the same room) are destroyed in a fire?

 

I was going to ask that question after having got a list for the main server going down on its own. If both servers go down in a fire a) I would have far more time to get things back up and running while the rooms / school are being fixed etc and B) I would have less options to start with, as I would only have the backups and no replication servers.

 

Any help on the process outlined in my OP would be great.

Posted
Anyone have any advice? Still not sure if I should Seize the FSMO roles and install OS manually or restore the entire server from backup. Would rather not work it out after an issue so any help really welcomed.
Posted

I would restore from backup, simply because the second server doesn't do that much. May work out significantly quicker too. You'd need to take the second DC offline if you did this, otherwise you could cause some issues with the directory (restoring to a live AD from a backup is bad).

 

As for fire - I have a physically separate room/building with an iSCSI target NAS in it - this holds all the bare metal restores for the servers that need it. You could also invest in a small, cheap server to replicate things in the same place (if possible).

  • Thanks 1
Posted (edited)

Thanks @3s-gtech.

 

So basically

Switch off DC2

On DC1 - Boot from Windows 2008r2 DVD and choose to restore the system using backed up bare metal system image.

 

Q1. is it really that simple?

Q2 what do I do with DC2 after the restore? Do I turn it back on or wipe it and re-build?

 

EDITED.

Edited by edutech4schools
Posted

That's what I'd do with your current system. Once you're 100% happy that everything is running okay again, you can wipe and rebuild DC2. This does mean that DC2 isn't that much use to you - you can increase its usefulness if you change your current setup a bit.

 

If you replicate as many functions as possible onto it, like DNS, DHCP, maybe even DFS-R, then you would have a live replica running which could take over if DC1 failed (once you transfer FSMO roles). Even without the file shares on it, you could have these backed up ready to restore onto a fixed DC1. All it would take is upping the spec (if necessary) of DC2 to handle this. Replicated DHCP needs 2012 or newer too.

Posted (edited)

DC2 does have DNS and AD replicating.

The second sever (DC2) is just a Global catalog DC with nothing else.
For it to be a domain controller (DC) it has to have DNS and AD. I choose to only have the GC FSMO role installed as its possible to seize the others if needed. Sorry if this was not clear in my OP but I did have a really bad cold when I created it.

 

Can we start again?

Edited by edutech4schools
Posted

DCs don't need to have DNS, but it makes sense depending on how many you have (my child domain has two DCs, neither hosts DNS). Obviously a DC has AD - they're synonymous.

 

What other roles on DC1 could you look at replicating to save yourself time and stress? You may be able to set it up as a print server. How about your MIS if you have one?

Posted

So what would be the best way to restore DC1 now you know DC2 has DNS AD etc?

 

Would it be to do a full bare metal restore on DC1? or install OS on DC1 and seize roles etc?

 

You may be able to set it up as a print server.

Good idea. Something I will look into. Thanks

Posted
Knowing that; I would rebuild the server, promote it to be a DC, restore the non-replicated parts such as MIS data (if needed), files and storage etc from backup, then test, test, test.
  • Thanks 1
Posted

My school district has seven buildings spread across a small city connected by private fiber. Using old equipment and some eBay purchases I setup an offsite ark in one of the other buildings that consists of two Dell Power Edge 2950s and an HP DL380G5 with an external MSA60 storage shelf. The Power Edges are running free ESXi 5.5 and the HP is running Windows Server 2012R2. The virtual servers are hosting two spare domain controller VMs and the storage box has 12TB of disk to serve as an offsite repository for Veeam. All the equipment is spread across two rack mount APC UPS units with monitoring cards. The UPS units themselves are pulls from elsewhere in the district with new batteries. Storage volumes on all the equipment are RAID5 with each having a dedicated hot spare.

 

This setup isn't really a warm site as I wouldn't be able to bring up all of our services, but it's all backed up and done nightly. The Dells were bought off of eBay for a lousy $250 a piece and came with dual quads, 16GB of RAM and remote management cards. The UPS batteries are only $25 a piece and we needed 16 of them.

 

Reliable offsite backup doesn't have to break the bank if you're willing to shop around.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...