March 25, 2010 at 7:53 am
Hello,
I have a question related to a mirroring issue I was deeply involved in a very very critical moment (the yard was full with trucks waiting for their documents to go to the stores)...
In that moment, I solved the problem as described below, but I know it wasn't the proper way to do it, but only for that desperate moment.
Can you help me with some ideas?
There were a lot of activities when the mirroring process was down and the both databases (both dbs became mirror, restoring) remained in "restoring" mode.
The solution was the easiest possible I could find in that moment:
1. Stop db activity.
2. Remove mirroring
3. Backups for both db (principal and mirror).
4. The biggest .bak was restored on the production
5. Mirroring reconfiguration.
6. Start db activity.
Another way to solve this better?
Thanks for replying,
Andrea
Wish you good ideas! 🙂
Andreea
March 25, 2010 at 9:15 am
First question, why was mirroring down?
March 25, 2010 at 9:31 am
I don't know. At that moment (I think an year has passed 🙁 ), I've looked in all the logs and it seemed to be a networking issue.
The witness had no major error and on all 3 servers, all the processes were up and unstopped.
Wish you good ideas! 🙂
Andreea
March 25, 2010 at 10:47 pm
I'm not positive because I have never tried to simulate it, but if none of the servers in a mirror-witness setup can see each other then there can be no quorum - you can't write to any. It's late and I'm tired, but I think you lose access to all of them; Microsoft recommend dropping the witness if the mirror is going to be offline for any length of time.
Matt.
March 26, 2010 at 12:57 am
Ok, but in order to restart the activity (very fast, without knowing the reasons of the failure, just because the business needs to go further), all the steps described and the method used is really the worst one I've could ever done?
Wish you good ideas! 🙂
Andreea
March 26, 2010 at 7:59 am
No, it's a valid approach in terms of taking backups and then restoring one, but your selection of the backup to restore might have been wrong. It would have been quicker / easier / more appropriate to determine which database was the principal and then restore it with norecovery to make it operational again.
March 26, 2010 at 8:52 am
Please tell me how could I know which was the principal?
Wish you good ideas! 🙂
Andreea
March 26, 2010 at 12:40 pm
You know which one was supposed to be the principal, you just have to confirm that it didn't failover - that event would be logged in the SQL logs if it occurred (or ideally you would have alerts set up to notify you of any mirroring events).
March 29, 2010 at 12:34 am
In fact, it did failover, that's why that database was in "Mirror, restoring" mode. But both of them were in "Mirror, restoring", and the only event in log was on both servers for the databases in mirroring... that they lost their connection with each other and then....total blackout.
Wish you good ideas! 🙂
Andreea
March 29, 2010 at 12:52 am
Viewing 10 posts - 1 through 9 (of 9 total)
You must be logged in to reply to this topic. Login to reply