Disaster Recovery Plan - Hardware/Software/Network

[Vegito]

Anyone actually bulit a real time disaster recovery site ? ie

I got 4 servers, 1 NAS, 1 SAN coming, 8 routers, 8 T1, 3 switches, 15 workstation.

It's a small shop but most of the data on the server are replicated to the NAS (160GB)

If our site floods, I can kiss everything good bye.

I can recover everything from Tapes or the NAS but crap, it could take days to do a full recovery ie

Get new computers with tape drive, get network switches, routers, etc.
Recover everything and test and then it'll be ready...

but that can take days... any of you guys done something similar or have ideas ? Off site disaster centers (of ur own) are expensive...

 

[Garion]

Never actually built one, but we did some studies. Some questions:

Are your 8 T1's voice or data? If they are data are you running point-to-point or frame relay? If you're running frame relay it's a lot easier - Just setup PVC's to your hot site. If they are point-to-point you're pretty much hosed for connectivity.

Do all 4 of your servers need to be replicated and brought on-line at the hot site or just a few critical ones?

Do you need to have ANY users beyond a few admin workstations or do you just need the servers and their data alive?

Cisco now makes a SAN-to-IP gateway that you could use to synchronize your updated SAN data across the network to your hot site.

Depending on your application, you might not need identical hardware. If you've got 160GB of space you could build a server with one or two of the new 180GB hard drives and have enough storage to survive.

You can either put this stuff at another site within your company, at a specialized DR site (There's lots of them out there) or at an outsourced data center like Exodus.

Good luck!

- G

 

[Vegito]

8 Data T1, 1 Voice T1, they're pretty much PTP T1s

Not all 4 server need to be up but the servers are

1 - SQL - DB - Main - Backs up on tape and NAS - Easy to recover
2 - Exchange 2K - Mail - Backs up on Tape - hard a$$ to restore to computer - hard hard hard !
3 - SQL - DB - Secondary Data - Backs up on tape ans NAS - Easy to recover
4 - File server & W2K Active Directory - Multimaster = easy, single = hard

The 160GB is a NAS to backup the above server

My plan was to pull the 160GB and bring up everything on another server but heck, it's hard to do a full recovery offsite..

Most of the stuff are already on another site...

This stuff cost too much...

 

[pissedoffwookie]

never done it either in the complete sense your looking for but have done some of it in pieces, i have a couple questions though

you mention that most of your data is in another location, does that location have an alternate access point? a small modem pool so critical users can emergency RAS into the alternate location may be an option.

also, although exchange and SQL aren't my area of expertise, it is my understanding that you can add a drive to bot the exchange and SQL server and only put the transaction logs on them, so that coupled with the tapes you have up to the second restorability

 

[Vegito]

Nope, secondary site isn't fully setup yet.. it's just a office with 2 server, 1 AD master to HQ and 1 IP through NATTing, not enough bandwidth to do real time transactions.. i tried, it took 3 hours to backup about 500 megs of file... need to get faster stuff

SQL is easy to backup & recover, I haven't really tried a full exchange restore to another computer in a different domain

 

[dotPablo]

Pricey, but depending on your risk assessment might be worth checking into;

http://www.comnews.com/cgi-bin/arttop.asp?Page=c0601backing.htm

 

[Marqui]

<< 8 Data T1, 1 Voice T1, they're pretty much PTP T1s

Not all 4 server need to be up but the servers are

1 - SQL - DB - Main - Backs up on tape and NAS - Easy to recover
2 - Exchange 2K - Mail - Backs up on Tape - hard a$$ to restore to computer - hard hard hard !
3 - SQL - DB - Secondary Data - Backs up on tape ans NAS - Easy to recover
4 - File server &amp; W2K Active Directory - Multimaster = easy, single = hard

The 160GB is a NAS to backup the above server

My plan was to pull the 160GB and bring up everything on another server but heck, it's hard to do a full recovery offsite..

Most of the stuff are already on another site...

This stuff cost too much... >>



Disaster Recovery is quite expensive any way you look at it. It's a matter of HA (High Availability), and if your company wants 99.999% uptime. There are plenty of options, but they are all pretty spendy. If uptime is a factor, you will need some type of i/o failover (pretty common on must clusters), and application failover and/or application resume. I know Xiotech has a pretty decent application called SAN Links Replicator. It may not be the right fit for the size of your company, but there may be smaller solutions available.