September 27, 2007 at 4:24 am
Hi All
We all know that the biggest thing that can happen is when your database goes down and you have to recover it.
Do some of you have a step-by-step guideline on how to do this?
Thank you in advance, your help would be much appreciated. 😉
September 27, 2007 at 12:04 pm
Not sure what you mean by "Step by Step". Each situation is unique.
In the event of a server crash then more than one database is involved and probably requires reinstalling the server instance, recovering the master database and all the active databases. If just a single database gets corrupted then it depends on the recovery model (FuLL or Simple) and having a good backup. BOL has pretty good instructions for how to do backups and restores.
For me personally I have a Notebook for each server I manage that is the "Disaster Recovery Plan". The notebook contains documentation about all the software on the server, current configuration information, license information and where needed step by step "reinstallation" instructions (this is often just the printed material provided by the software manufacture with appropriate notes where needed), etc. This includes the SQL Server software (Version and current patch level) a list of all the active databases to include associated files and hard drive locations. Additionally this document lists how and "When" (schedule) the backups are performed and all the steps and commands that would be required for a complete recovery of the operational system on new hardware in the event of a complete loss of the server. These instructions can be used for partial recovery obviously. This type of document is a lot of work and needs to be reviewed monthly and kept up-to-date (especially when new software is installed or major software upgrades occur). Additionally it is best to "TEST" the document periodically. I've always been with organizations that have had sufficient resources to have real Development/Test servers which are available for testing the DRP (i.e. blowing the box away and starting from scratch), which I normally do every six months. (I just changed organizations and this is the first where they did not have any development/test hardware/software. I've recently rectified all of that and am in the process of writting the DRP right now!). Additionally I keep a copy of all the software installation media in the back of the notebook (this way it's readily available when needed), the originals, if any are kept where ever the organization normally keeps it's software (which often is not well defined!).
I would recommend that you do something similar if possible (I know not everyone has the luxury of having sufficient equipment to duplicate a production environment).
All the above (while long winded I know) was basically to say that nothing beats your own Step-by-Step instructions for recovering from a problem. If you need help with different steps of the recovery process then just post information about your configuration and the type of backups you have and the problem (i.e. master corrupted is a lot different than a user database corrupted) and I'm sure myself or someone else here would be happy to post the commands/steps to recover that specific database.
James.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply