I have a database running fine, however, it is purely stock configuration. I
would like to fine tune it for better performance and to control memory
I have seen people refer to restricting the number of back end processes,
and fine tuning maximum memory consumption as well as other items.
I've read the man pages, and a few options look promising, such as -B on
postmaster and -S on postgres. However, I can't find anything that explains
what exactly those options will do, and the ramifications of changing them
away from defaults.
I also can't find any mention of restricting the number of simultaneous
connections. Or, restricting the number of back end processes that can be
The database will have a lot of information, but has no need to run more
than maybe 4 to 12 backend processes. I would estimate that it wouldn't need
more than 6 unless special circumstances arose.
This database is going to sit behind a web based front end. There won't be a
lot of simultaneous queries on the system, in normal operations. The web
will be used as a query tool. Data modifications and updates will probably
be via Access and ODBC, but that may change to some web based modifications
and updates at a later time.
Does anyone have any pointers to fine tuning memory and performance details
of Postgres? I would really like to find something that documents not only
what things do, but what ramifications of certain operations are. I want to
understand the performance tuning, not just copy someone's prewritten
Gregory S. Youngblood
pgsql-admin by date
|Next:||From: Ricardo AG Almeida||Date: 1998-10-02 19:41:10|
|Previous:||From: Melanie Renee Nelson||Date: 1998-10-02 18:59:06|
|Subject: compiling on SGIs|
pgsql-general by date
|Next:||From: Taral||Date: 1998-10-02 21:12:19|
|Subject: RE: [HACKERS] RE: [GENERAL] Long update query ? (also Re: [GENERAL] CNF vs. DNF)|
|Previous:||From: Bruce Momjian||Date: 1998-10-02 17:58:25|
|Subject: Re: [HACKERS] RE: [GENERAL] Long update query ? (also Re: [GENERAL] CNF vs. DNF)|