Question DB crashed

Hello Guyz,

I'm just out of my first db crashed. We can't found the issue.
But we discovered after that our two last backup before crash were corrupted.
Do you know if there is a way after restore a backup to test if the process was successfull or not ? Or even without restoring the backup in a dummy environnement?

Thank you for your help.
Best regards,

BobyIsProgress
 

TomBascom

Curmudgeon
You answered your own question... PROREST the backup on your test system.

There is also prorest -vf but that requires that the source database be offline for the backup and the validation. That is usually not practical.

You could use -red on the backup and -vp on the restore to improve resiliency. That is also not usually very practical.

More importantly - make sure that after-imaging is properly configured and enabled and keep a deep history of backups and archived ai logs. That way you can overcome corrupt backups without losing data by going to the last good backup and rolling forward.

You might also want to implement OpenEdge Replication.
 
Thank you @TomBascom for your quick answer as always :)
After the PROREST, is there a way to know that the Dumy database has restart properly and logged it?

I don't know if a proenv command is abled to do it or a simple connexion test would do the trick ? I need to know if the Dumy database is running smoothly afte the PROREST .
 

TomBascom

Curmudgeon
It depends on how much reassurance you need and what sorts of problems you are trying to protect against.

A couple of simple sanity tests:

1) start a single user session with "pro" and then quit. That confirms that the database was restored in a minimally functional manner.

2) run "proutil dbname -C dbanalys" that will read the entire database and provide a level of comfort that it is physically consistent (you might still have logical errors though)

3) start a server and run an application test suite that exercises everything (ok, maybe that isn't "simple")

Keep in mind that opening the database (starting a server, connecting with a session, running utilities) without the -RO option means that you will not be able to roll-forward after image files. So if a roll-forward is part of your recovery you need to do that first.
 
Thank you @TomBascom I will keep in mind what you said.

The behavior I had with my crashed backup was after the PROREST the server wasn't able to start at all. So the first solution would be already good.
I will see what I can get from a dbanalys option but could be a good way to be sure of the sanity of the backup.

I will get back you on what we decide and the code I will use to do it
 
Top