Re: Load spikes on 8.1.11 - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Load spikes on 8.1.11
Date
Msg-id 9793.1216355319@sss.pgh.pa.us
Whole thread Raw
In response to Load spikes on 8.1.11  ("Gurjeet Singh" <singh.gurjeet@gmail.com>)
Responses Re: Load spikes on 8.1.11  ("Gurjeet Singh" <singh.gurjeet@gmail.com>)
List pgsql-hackers
"Gurjeet Singh" <singh.gurjeet@gmail.com> writes:
>     During these spikes, in the 'top' sessions we see the 'idle' PG
> processes consuming between 2 and 5 % CPU, and since the box has 8 CPUS (2
> sockets and each CPU is a quad core Intel Xeon processors) and somewhere
> around 200 Postgres processes, the load spikes to above 200; and it does
> this very sharply.

This looks like heavy contention for a spinlock.  You need to get a
higher-level analysis of what's happening before anyone can say much
more than that.

Note that 8.1 is pretty much ancient history as far as scalability to
8-core hardware goes.  You should probably consider updating to 8.3
before investing too much time in tracking down what's happening.
If you can still show the problem on 8.3 then there would be some
interest in fixing it ...
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Re: [COMMITTERS] pgsql: Add column storage type to psql \d+ display.
Next
From: "Gurjeet Singh"
Date:
Subject: Re: Load spikes on 8.1.11