I know that this type of information is probably buried in
documentation, but I would like an opinion based on 'experience', which
I'm sure there is quite a lot of that here.

I have a situation here where the database space has 'grown' to about 45
GB in use, with 65 GB available, but only 4 GB of disk space allocated
for JNL, CKP & DMP in total (not each! - i.e. all three sharing this 4
GB). They are not using the CKP area as its now too small, so
checkpoints are being re-directed to other storage area (outside of
ii_checkpoint), like you would do to tape. Journalling is frequently
causing disk space availablity to become a problem because of the size
of the databases being stored. So we need to consider re-locating the
ii_checkpoint, ii_journal, ii_dump to larger areas using alterdb, but
the question is, how much now?

What is the suggested disk space allocations, from experience, given
this scenario, or for any basic setup?

Thanks for any suggestion or views.

- David.

--
-----------------------------------------
David G. Brooks
Vision Systems Software
(P.O. Box 804, Bridgetown)
Omega Building - Suite 5
Dayrells Court Business Centre
Dayrells Road
Christ Church
BARBADOS
West Indies

URL: http://www.VisionSystemsSoftware.com/

Tel: (246) 431-0193 - Work
(246) 436-1417 - Home

Fax: (246) 431-0192 - Work

Hello David,

Checkpoint location :
If you want to start checkpointing to disk, then you need at least the
size of your database location.
If you want to keep x consecutive checkpoints on disk, then you need x
times your database size.
Practically, I keep one checkpoint to disk and previous ones are
backup up to tape. I also keep an extra diskspace available for about
2 checkpoints. Your tape drive can be malfunctioning and a replacement
can last 2 or 3 days. So, therefore I try to have about 3 times the
database size for the checkpoint location , just in case I have to
checkpoint to disk for 3 days.

Journal location :
Having half the database location is fine. Actually depends on the
"rotation" of your data.

Dump location :
Same as journal location.

Don't forget the work location :
the biggest of either (twice your biggest table), or eiter
(the size needed for sorting your average intermediairy query result x
number of (concurrent sessions in average)). The latter can not be
calculated and should be based on observation.
I use minimum twice the biggest table.

Regards,

Franky Leeuwerck
Advanced Projects & Products
http://www.ap2.be