Beefy Boxes and Bandwidth Generously Provided by pair Networks Frank
laziness, impatience, and hubris

Re(3): Sysadmin Aspects of DBMS

by Arguile (Hermit)
on Nov 24, 2001 at 04:26 UTC ( #127198=note: print w/ replies, xml ) Need Help??

in reply to Re: Re: Sysadmin Aspects of DBMS
in thread Sysadmin Aspects of DBMS

"[R]equiring server installations..." I can only imagine you're used to working with flatfiles from that line ;). I'm mainly going to mirror Thraxil here, I've worked with Oracle, DB2, and PostgreSQL extensively and PostgreSQL does have a rather light maintenance load.

How much work needs to go in will of course depend on your load, dataset, and the equipment you're running on. For many the defaults will be sufficient. Personally, even for light loading, I like spending a bit of extra time on install and tweaking everything to the hardware. Obviously you can do quite a bit partitioning, managing tablespace, etc. but

From the sound of your data/load setting up a couple cron jobs and applying the occasional patch will probably be all the sysadmin work you'll need to invest in it. VACUUM and VACUUM ANALYSE (or the command line wrapper vacuumdb as noted by Thraxil) and some backup cron jobs will probably be the extent of your involvement.

If you're going to be the DBA as well then an occasional audit of expensive queries (using EXPLAIN) and indices maintenance will often prove fruitful; though not required unless you're experiencing poor performance. Again, this is based on what you've said of your application.

To get a peak at the O'Reilly book before it's published (or to preorder), Command Prompt (the writers) have it online here. Bruce Momjian's PostgreSQL: Introduction and Concepts is also available online. And you can of course always read the official Administrator's Guide.

P.S. You may want to ask the editors to put an [OT] in the title of the post.

Comment on Re(3): Sysadmin Aspects of DBMS
Re: Re: Re: Re: Sysadmin Aspects of DBMS
by jlongino (Parson) on Nov 24, 2001 at 04:49 UTC
    Actually that does seem a silly question in retrospect if talking about DBMS. And you're somewhat correct in assuming most of my Perl/database work has been with flat files, but I have done quite a bit of work with pure MS Access (regrettably). I was thinking of cases like Access where you can do what's necessary via ODBC/DBI without needing MS Access installed (provided your database file is already created).

    Thanks for the links.


Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://127198]
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others cooling their heels in the Monastery: (13)
As of 2014-04-25 08:53 GMT
Find Nodes?
    Voting Booth?

    April first is:

    Results (585 votes), past polls