Is it normal to have too many idle processes in postgresql cluster?
Hi all, I have a database cluster running Postgresql 9.6 on RHEL 6 servers.
I've noticed a lot of idle processes in the master database as opposed to
the slave database and from the looks of pg_activity, they're all idle
statements like "SELECT 1". Forums and StackOverflow seems to suggest its
nothing to worry about but
[user@db1.pr ~]$ ps -ef | grep postgres | grep idle | wc -l
213
Seems to get me worried. Should I be worried? Is this normal? If not, what
should I do?
--
Abdul Qoyyuum Bin Haji Abdul Kadir
HP No: +673 720 8043
On Mar 7, 2021, at 17:34, Abdul Qoyyuum <aqoyyuum@cardaccess.com.au> wrote:
Seems to get me worried. Should I be worried? Is this normal? If not, what should I do?
It's normal. Every connected client has an associated backend process, even if the client isn't doing anything. That's what the "idle" state means. You can consider using a pooler such as pgbouncer to reduce the number of connections the clients have open, although 200 open connections is nothing concerning.
--
-- Christophe Pettus
xof@thebuild.com
Thanks Christophe! I guess I had nothing to be worried about.
On Mon, Mar 8, 2021 at 9:46 AM Christophe Pettus <xof@thebuild.com> wrote:
On Mar 7, 2021, at 17:34, Abdul Qoyyuum <aqoyyuum@cardaccess.com.au>
wrote:
Seems to get me worried. Should I be worried? Is this normal? If not,
what should I do?
It's normal. Every connected client has an associated backend process,
even if the client isn't doing anything. That's what the "idle" state
means. You can consider using a pooler such as pgbouncer to reduce the
number of connections the clients have open, although 200 open connections
is nothing concerning.--
-- Christophe Pettus
xof@thebuild.com
--
Abdul Qoyyuum Bin Haji Abdul Kadir
HP No: +673 720 8043