how to deter. witch AI files for roll forward

[Moussa]

Hi,

to roll foward I first restore the back-up from the night before and then do a
"for AIFILE in /bla/db.ai*; do ; ${DLC}/bin/rfutil DB -C roll forward -a $AIFILE"

It works, but as we save AI files for 4 days, the loop walks all ai files, the most of them are not suitable for roll forward.

My question is, how do I determine witch AI files I need, so i can filter these?



ty.

 

[TomBascom]

It depends.

One simple method is to use a naming convention for your saved ai extents that makes it obvious. This is especially easy if you happen to backup at midnight -- just use a date based naming convention. Then script accordingly.

Another is to take note of the current ai extent when you create the backup and to log that information somewhere where you can retrieve it.

Or you could take a peek into the database .lg file

Of course there are lots of other methods that you could use too. But no matter what the method you probably want to modify your roll forward script to accept a starting extent parameter and an (optional) ending extent or # of extents to roll forward.

Personally I'm usually pretty lazy -- if I know that the backup was done at 3am and that extent switches are every 15 minutes I'll probably start the roll forward with the extent that was saved at 2:45. I support this bad habit by either ensuring that the time stamps on the saved extents don't get changed or by renaming them to reflect the date and time that they were saved.

 

[ron]

Yep, all of those methods work, of course.

I prefer to move the extents into a different directory. I guess I like things in "tidy boxes"!

 

[Moussa]

Hi Tom,

thanks for your input.

Because we start switching and copying AI extents at 6h in the morning of the next day, I'll go with a find command to provide this files. Hope it all works out.

Regards, Moussa.