Re: Journalling All Files

Journalling *ALLUSR/*ALL *ALL is possible-- we do it (at least with our main application libraries) so that we can replicate our data to our backup machine using MIMIX. MIMIX uses remote journals to pass the data to the 'other' machine.

We only keep a rolling 48 hour supply of journal receivers on the system-- MIMIX handles the cleanup.

Keeping everything would be huge! We add about 2 gig of new data to our main application files every day.

What we do is extract the journal entries for certain key files-- master files, some transaction files, some audit files, and other files that programmers and management can use to track down who did what to whom.

We retrieve the journal transactions for these files, the store them in a formatted file-- the journal data, followed by the data fields in the original file. We don't keep the original journals.

We dump these formatted files to tape, and keep a rolling week of transactions on line for programmers and problem solvers.

Keeping everything will generate tons of data. Unless you really need all of that data, it's better to examine your processes and find out which files will give you the best snapshot of what's happened, without strangling the system.

--Paul E Musselman
PaulMmn@xxxxxxxxxxxxxxxxxxxx