Re: Terminate the idle sessions - Mailing list pgsql-hackers

From Li Japin
Subject Re: Terminate the idle sessions
Date
Msg-id 2F900B6E-5F2C-4D38-AA50-00E2C6C0FD46@hotmail.com
Whole thread Raw
In response to Re: Terminate the idle sessions  ("David G. Johnston" <david.g.johnston@gmail.com>)
Responses RE: Terminate the idle sessions  ("kuroda.hayato@fujitsu.com" <kuroda.hayato@fujitsu.com>)
List pgsql-hackers

On Nov 17, 2020, at 10:53 AM, David G. Johnston <david.g.johnston@gmail.com> wrote:

On Monday, November 16, 2020, Li Japin <japinli@hotmail.com> wrote:

<note>
Consider setting this for specific users instead of as a server default.  Client connections managed by connection poolers, or initiated indirectly like those by a remote postgres_fdw  using server, should probably be excluded from this timeout.

        <note>
         <para>
-         This parameter should be set to zero if you use postgres_fdw or some
-         connection-pooling software, because connections might be closed unexpectedly.
+         This parameter should be set to zero if you use some connection-pooling software, or
+         PostgreSQL servers used by postgres_fdw, because connections might be closed unexpectedly.
         </para>
        </note>


Prefer mine, “or pg servers used by postgres_fdw”, doesn’t flow.
 
Could you please explain how the idle-in-transaction interfere the long-running stability?

From the docs (next section):

This allows any locks held by that session to be released and the connection slot to be reused; it also allows tuples visible only to this transaction to be vacuumed. See Section 24.1 for more details about this.

Thanks David! Attached.

--
Best regards
Japin Li
Attachment

pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Tab complete for CREATE OR REPLACE TRIGGER statement
Next
From: "lchch1990@sina.cn"
Date:
Subject: Re: Add statistics to pg_stat_wal view for wal related parameter tuning