ADSM-L

DataBase File Spaces and Selective Backups

2015-10-04 17:12:14
Subject: DataBase File Spaces and Selective Backups
From: Seay, Paul [mailto:seay_pd AT NAPTHEON DOT COM]
To: ADSM-L AT VM.MARIST DOT EDU
Right now we do a lot of our database backups offline without a TDP.  Many
of these will stay that way because they are very small databases and
setting up the TDP is just too much work and cost.  However, we are
concerned about what this means from a filespace and backups point of view.

First of all, because we never perform an incremental against these
filespaces we have no filespace statistics.  Second, it is very difficult to
monitor selective backups because you have to search the backups table for
the entire filespace name and pick the last backup from it to get something
similar.  We have excepted these restrictions but we started thinking about
what does this mean if an incremental accidentally got run so we are
considering another approach.

We believe we should be doing an incremental on every database filespace no
matter what except do an exclude of every file, not directory, but file, ie.
do not use the exclude.dir.  However, we are concerned that this may expire
all the selective backups.  Before we tried it we thought we would ask what
everyone else is doing.

The other option would be to do a selective and then an incremental which
would backup nothing because it did not change.  There are only 150 files
max in these filespaces anyway, so it is no big deal.

You may wonder why we want to do it this way.  Our plan is to segregate the
database storage pools here so we can eliminate the reclamation churn.
Basically, everything would expire all at once and the tape would be
immediately empty.  However, to make that happen, we must backup everything
every time.  99% of the data files change anyway so incremental would do
nothing.

Paul D. Seay, Jr.
Technical Specialist
Naptheon, INC
757-688-8180