I think the log file would be too big to post since the program runs for hours............however you are right...i did some analysis and found that the program runs slwo when another loader is running. If the loader has run for 2-3 hours.......the pro
gram has run extremely slow at that time....and hence taken longer to complete.
However the loader acts in the main table and the program works on the partitions of the main table.
Iam not sure what is the exact problem here. Could throw me some insight on the internal working of potsgres?
Sam
On 7/15/08, Vishal Arora <aroravishal22@hotmail.com> wrote:Hi,
Can you show your event log file?
There might be that at different point of time the load on DB is different so it takes some time more than what is expected.
-Vishal
Date: Tue, 15 Jul 2008 13:54:00 -0400
From:
sam.mahindrakar@gmail.comTo:
pgsql-admin@postgresql.org Subject: [ADMIN] Jobs using pgagent
Hi....
Iam using postgres since recently.....there are jobs scheduled using the pgagent. There are loaders and stuff that are scheduled using the pgagent. There is a new job that i have added............the weird part is that it runs for different amounts of time everyday. For example: it has run for 3hours,5 hours on some days and for 16hours on some days. It processing equally amounts of data on each of these days.
When i look at the log file............the RAISE NOTICE statements are not continous .........i.e. statements of other DB processes are printed and then the statements of the program are printed............seems like its running on and off. As far my knowledge goes pgagent can handle parallel jobs.
Though i see locks made by the program............iam not sure if thats the problem. Can someone give me a pointer on this?
Thanks
Sam
Check news, cricket, entertainment and astrology right from your mobile. Browse
http://m.msnindia.com from your GPRS mobile phone.
Try it now!