Discussion Locked

Disaster Recovery /Business Continuity

0 helpful
7 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

Hi ,

We are looking for some advice on a disaster recovery solution for a site with an NT 4.0 server (PDC,data), exchange 5.5 (on NT4.0 server). The site cant be down for more than an hour or two in case of a disaster . We currently have them backing up using Arcserve and a DAT40 which normally suffices for most of our clients which arent as critical. The site has 150 users and is growing so we need a solution to keep up with that.

Any advice appreciated

Thanks

7 replies

by scottsman

HI, if you are looking for a 99.9 uptime with room for 300 - 400 clients(future growth) like my IT dept is. We are implementing a 2 server Dell cluster with a disk array. What happens is the 2 servers share the disk array which decreases our pointsof failure to extended loss of Power or natural disaster/fire only. The array is redundant with multiple Hard Drives (ours is 5) in a RAID5 config with redundant power supplies as well. If one server goes down or is being maintained the other picks up the slack. (you can also enable load balancing to utilize both servers efficiently) Note that this solution does require a capital investment.

by cactus pete

You need redundancy everywhere. Depending on how much you really mean that you cannot be down for more than two hours, you need a failover to a colocated site. If you already have a second office, this may suffice.

There can be no set answer injust this forum - it is a large LARGE project you're talking about.

You'll also likely need to upgrade your platform, if you really want a fast turn around in case of a disaster, as NT does not have a long support life left.

I highly recommendyou consult with a local firm that can help you through this migration with an eye towards DR.

by sgt_shultz

you didn't mention BDC--? time to ditch nt4 and go to win2000 and active directory imho. other advice very good: redundancy everywhere. you aren't going to fix much in an hour or two. might take you that long just to get to site to respond. you wantauto-fail-over stuff. like clustering. getting help from outside IT good idea but make sure *you* are given understanding of how stuff works. exactly. there are implications EVERYWHERE so ask lots of questions. think of everything that can fail, figure out exactly how your will respond in each event. motherboard failure. ram failure. power supply failure. RAID drive failure. virus. power outage. isp outage. telco outage. fire. flood. tornado. quake.
get critcal backups and phone numbers, partslists etc off site. repeat: far far away. when quakes hit in CA, folks had fire safes which safely protected tapes etc (amazingly). story was, they weren't allowed back into site to get to the tapes. so dead in water anyway. and/or, dozers came in and scooped safes into landfill. urban myth maybe, but food for thought...
a good disaster recovery plan has to be revisited at least once a year. it is ongoing budget cost...forever more...

by johngreer

To do it right, there really is no simple answer.

You should really analyze the applications/business processes
and design "on paper" a BC solution for each,
based on each applications business needs.
At the very least you should ask each application/business process:

1) How long can it be down?
2) How much data can if afford to lose?

Once you answer that, we can design a solution for each.

Once the solutions are priced and justified individually,
I then look at each to see if they can be combined (hardware, software, telecommunications, etc.) in order to get the best ROI for the business.

If you'd like to talk more, feel free to e-mail me.

by tim.doyle

First, define disaster. It doesn't always mean "Office burns to the ground", it could just mean power outage. It could be a situation where your systems are safe, but the office is unoccupiable...like if a railcar full of chlorine ruptured on the tracks near the office. What you have to do is talk to your clients and figure out what contingencies they want you to help them plan for. Ask why they are doing this? Is it because they're a financial organization and they need to do better on their IT audit? Is it because they have had trouble in the past? Is it because their business is so critical that they really and truly cannot afford more than 2 hours lost time? Do they run 24x7x365?

I'm a member of my company's (bank) disaster recovery and business resumption planning and testing team, responsible for the IT side of things. If the last year of work has taught me anything its that you can never have a robust enough plan...there are stil things that will be lacking.

You need to do alot of talking to the department heads and find out what do they truly need to be up and running for 1-7 days. Longer than that and you can have pretty much anything fixed or replaced except a building.

by tim.doyle

I'm settingup my own "hot" site right now...each department will be allocated at least one permanent hot site PC and their critical data (GL stuff for example) will be copied to my offsite server daily. Jetdirect boxes are pre-programmed and all I need to do is snag a spare paser printer from upstairs and I'm up and running.

As noted above, the plan needs to be revised every year and then tested and revised again to fix any problems found during testing.

You have alot of work ahead of you...it will likely take months of planning and interviews with other staff members adn the cost can be very high depending on the solution you opt to put in place.

Good luck and don't forget to check back in with us.

by wlbowers

Dump nt and upgrade to Server 2003

You will need to run a cluster with one or more of the servers and it's raid box in a colocation. The task of restoring data from tape is going to take more than the down time maximum.

Setup up the servers with mirrored os hard drives. Set up one of these os drives in a hot swapable bay. Have a third hard drive in a second tray. Now rotate these removables. Well just ask your self what are you willing to loose.

Take this drive off site or store in a heavy fireproof safe.

Now have a tape backup in one or both locations.

Off site redundency is the technology being used today for critical operations.

The technology is there. It just depends on how safe you want to be.

Good Luck Lee

This discussion is closed to new replies.