Disaster Recovery Scenario

[antzzz]

Hi,

We're running SQL 2000 on a Windows 2000 server. Drive D is on EMC (75 GB) which contain the User Database data files (.mdf/.ldf); Database1 has total of 20 GB while Database2 has total of 50GB (leaving only 5GB free space). Drive E (50 GB) is also on EMC and is used to hold the Database backups. Backups are done 1 database at a time 'coz Drive E can only hold 1 database dump at a time. System Data files are on C: (local hard drive). Database backup files are backed up using Veritas Tape Backup Solution before they are erased from Drive E. It takes 3 1/2 hours to backup the 50GB Database1 backup to tape. 20GB Database2 takes about 1 hour to backup to tape.

Our Server admin will be installing new local hard drives on the system and he says that during installation, there might be a risk of corrupting the existing local hard drives. He says that if this happens, he will reinstall Windows and I have to reinstall SQL Server. He says that the EMC drive will stay intact and not be corrupted. This means that my SQL Data files (user databases) will be safe.

As the server is a very critical one, we're only allowed a downtime of 1 hour.

I have the following disaster recovery process if my EMC drive is still intact (without resorting to backups):

1) After I shutdown SQL Server, I copy the master and system data files to Drive E or D (EMC)
2) Let's pretend he messes up the local hard disk and reinstalls Windows 2000 server.
3) I then reinstall SQL Server and I shut it down.
4) I then replace the newly installed master and system data files with my copies saved on EMC.
5) I then bring up SQL Server.

OR

4) I keep the newly installed master and system data files and just reattach the user data files that are on EMC using sp_attach (i.e. without detaching them in the first place)
5) I then bring up SQL Server and just recreate the orphaned logins.

What I would like to know is whether any of my disaster recovery plans will actually work?

Any advice is appreciated.

Thanks

 

[SQLSister]

My first advice to you is to make a full backup immediately before any of this happens and to store it on a tape or a seperate server. Also script any DTS packages you have.




Questions about posting. See faq183-874

 

[mrdenny]

Yeah, I'm with SQLSister here. A Full backup is going to be a must. I'm not sure how adding new drives can screwup the current internal drives. If it does, he did something wrong.

I like your first idea better, as you don't have to recreate your jobs, dts packages and logins.

Denny

--Anything is possible. All it takes is a little research. (Me)