I am using PostgreSQL in an embedded system which has only 32 or 64 MB RAM
(run on PPC 266 MHz or ARM 266MHz CPU). I have a table to keep downlaod
tasks. There is a daemon keep looking up the table and fork a new process to
download data from internet.
. Check the table every 5 seconds
. Fork a download process to download if there is new task
Downlaod process (there are 5 download process max):
. Update the download rate and downloaded size every 3 seconds.
At begining, everything just fine. The speed is good. But after 24 hours,
the speed to access database become very very slow. Even I stop all
processes, restart PostgreSQL and use psql to select data, this speed is
still very very slow (a SQL command takes more than 2 seconds). It is a
small table. There are only 8 records in the table.
The only way to solve it is remove all database, run initdb, create new
database and insert new records. I tried to run vacummdb but still very
Any idea to make it faster?
Here is the table schema:
create table download_queue (
CONSTRAINT download_queue_pkey PRIMARY KEY(task_id)
CREATE INDEX download_queue_user_index ON download_queue USING BTREE
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
pgsql-performance by date
|Next:||From: Csaba Nagy||Date: 2005-11-19 11:12:12|
|Subject: Re: VERY slow after many updates|
|Previous:||From: Jaime Casanova||Date: 2005-11-19 04:55:26|
|Subject: Re: What is the max number of database I can create in an instance of pgsql?|