General discussion

Locked

Disaster recovery: how do you do it?

By Net Designer ·
I am currently trying to come up with Disaster Recovery Plan for our 4 offices (one main and three small sattelite). We are a Novell shop and my boss doesn't really have much of a budget to do it. So, to go on the 'budget', I came up with idea to place 'backup' servers at each office (just an old PC with large enough HDD), do a plain xcopy from main server to 'backup', which will also hold a copy of NDS database. That's for the case of main server failure. For event of site's failure, I will becopying data to one of the off-site 'backup' servers. So, if say Site1 goes down, we go and use Site2's 'backup' server that would have our data on it, as well as NDS DB for authentication.
Do you think this plan would cover all I need, as my boss is trying to push me to also have a tape backup of the 'backup' server and rotate those tapes between sites. He says that virus could destroy both servers. I think he's overcomplicating simple process. What do you think? Besides, I am trying to get rid of tape backups all together, as I don't have manpower to support them in satellite offices and they are not very reliable either.
Thanking everyone for the input you may provide.

This conversation is currently closed to new comments.

4 total posts (Page 1 of 1)  
| Thread display: Collapse - | Expand +

All Comments

Collapse -

Some input...

by Amadeus Hack In reply to Disaster recovery: how do ...

I am going thru that @ my company right now....if you do a search on www.google.com and type in "DRP", you can get some really good information....

Here are some things you need to look @ when planning DRP....

Acceptable Downtime: What is an acceptable amount of downtime for your company, with that also goes what is your acceptable downtime for each of your offices??? We have 12 locations and figured out that 4 of them MUST be up ASAP, the other can go for a few days of downtime and therest could easily go for a week or more.

Connectivity: How do your other offices connect?? Do you have Internet VPN's, Frame-Relay, Dial-up? Now we have Frame-Relay/ATM, and we are a hub and spoke network, (all of our sites must connect with our main office to get to the internet, sql, external email, etc...) We are placing internet connections at our must be up offices and can switch over to VPN if the frame goes down (you can work out deals with an ISP to get internet connectivity at other offices really cheap, you pay for the install, and a small monthly fee but can have it set up so you only pay out the big bucks if you have to use them). Or you can use business DSL, it works great!!! Figure out what offices need some sort of connectivity whether it is internet or just back to a main office.

See Next Post....

Collapse -

Some input part 2....

by Amadeus Hack In reply to Disaster recovery: how do ...

Backups: You MUST have some sort of backups, if not the operating system, at LEAST the data on the servers. Even if you back it up to hardware you will be better off. I have never know an IT person who didn't get canned if a server went down and they didn't at least have the user/company data backed up. For your safety (and sanity!!!) please do backups. Most backup software, you can automate so you don't need the manpower...You can even set up a cd burner with a rewritable cd to do "important" files only every night and automate it so you never have to change the cd.....

Equipment: Do you have enough equipment to get the company back to some sort of "production" mode? If you don't, can you "load" up the programs/data onto anotherserver that is already running and run it from there. Do you have a way for users/management to work if you have a disaster that takes out a "work area"?? You could possibly set up a VPN server, (you can get software for FREE for some sort of VPN server) that way at least management could work from home or say a Kinkos or cyber cafe.

If you at least cover these item, you have the basic minimums for some sort of DRP. I understand that more and more companies are on a budget, but if you have a site go down and you can't do business, how can they make any money at all.

See Next Post....

Collapse -

Some input part 3.....

by Amadeus Hack In reply to Disaster recovery: how do ...

I have seen it more than once where companies want to piece together a DRP, and when anything happens they wished they would have covered those minimums I explained above. Also from experience you are the first person mgmt will look at if anything happens and you will get grief...or loose your job. Do yourself a favor and anything you may or may not implement, you need to bring that before company mgmt (all of them, not just your boss) and explain the implications of not having the pieces in place and make sure they understand the consequences of not having a minimum DRP (you may trust/like your boss, but when something like this happens, if you are the one who was supposed to do it, it will be YOUR responsibility). Please email me.... amadeus.hack@verizon.net if you have any questions or if you want to give me more details on how your network is set up, I will be more than happy to help you figure out what might be the least expensive solution for you....Thanx!!!

Collapse -

Thanks for input

by Net Designer In reply to Some input part 3.....

I am pretty much covering all basics: onsite backup to 'hot standby', offsite backup to offsite's 'hot standby'. Backups are done with plain xcopy from win98 PC every hour, so the downtime is very minimal. The offsite xcopy is done once a day, so ifsite goes down, the most they'd loose is one day's work. The 'hot standbys' also hold the copy of NDS Replica for that site's context (basically a user IDs database for that site). Printing is also a redundant, we have main production run NDPS and the 'hot standby' have queue-based printing set up. All users have both type of printers installed, so if main is out, they could use queue-based that's already there for them.
Additionally, we have tape backups that backup 'hot standbys' at night after all xcopies are done, without affecting main production servers. Our WAN is an SDSL-based VPN going through Internet via tunnels, each site has two different ISP providing DSL service, all DSL routers are tunnelled to each other (are set up as a full mesh), and are RIP-enabled, giving 100% redundancy. Since the 'hot standbys' don't have to be top of the line servers, we went with plain Dell PE with no RAID, just HDD big enough to hold all the data we backup. The whole deal cost us under 10K in hardware and software. I think it's pretty good, considering what's covered.
Many thanks to you for your ideas.

Back to IT Employment Forum
4 total posts (Page 1 of 1)  

Related Discussions

Related Forums