BUG #3808: Connections stays open in state CLOSE_WAIT

The following bug has been logged online:

Bug reference: 3808
Logged by: Daniel Migowski
Email address: dmigowski(at)ikoffice(dot)de
PostgreSQL version: 8.1.4
Operating system: Windows Server 2003
Description: Connections stays open in state CLOSE_WAIT
Details:

Hello dear bug report readers,

I am using PostgreSQL 8.1.4 on Windows 2003 Server and do my backups with
the pg_dump in the same package.

From time to time the server side of the connection remains open with
netstat showing the connection in state CLOSE_WAIT. The pg_dump process at
this time has succesfully done its job and is gone long, but the connection
stays open.

I notices this, because, together with the connection, many AccessShare
locks on different tables and one ExclusiveLock remains open.

It is possible to cancel the backend process from PGAdmin, but after
grinding the server to a halt with the most minor update script (it was a
DROP NOT NULL), I really get annoyed about this bug.

Even more problematic is, that the behaviour cannot be reproduced (i tried
for hours), it just occurs occacionally.

I noticed the changelog of version 8.1.10 (Magnus: Windows Socket
Improvements). Will they fix this bug?

Do you need more information to fix this bug?

Thank you very much for taking your time to cover this in advance, and with
best regards,

Daniel Migowski

Daniel Migowski wrote:

> I noticed the changelog of version 8.1.10 (Magnus: Windows Socket
> Improvements). Will they fix this bug?

Upgrade and tell us.

We don't put out bugfix releases just for the sake of wasting bandwidth.
If you're not running the latest version from your stable branch, IMHO
your bug reports should be ignored.

--
Alvaro Herrera http://www.amazon.com/gp/registry/CTMLCN8V17R4
"In a specialized industrial society, it would be a disaster
to have kids running around loose." (Paul Graham)

Alvaro Herrera wrote:
> Daniel Migowski wrote:
>
>> I noticed the changelog of version 8.1.10 (Magnus: Windows Socket
>> Improvements). Will they fix this bug?
>
> Upgrade and tell us.
>
> We don't put out bugfix releases just for the sake of wasting bandwidth.
> If you're not running the latest version from your stable branch, IMHO
> your bug reports should be ignored.
>

On the bug reporting form it says:

Make sure you are running the latest available minor release for your
major version before reporting a bug. The current list of supported
versions is 8.2.5, 8.1.10, 8.0.14, 7.4.18.

Regards, Dave.

Dave Page wrote:
> Alvaro Herrera wrote:
> > Daniel Migowski wrote:
> >
> >> I noticed the changelog of version 8.1.10 (Magnus: Windows Socket
> >> Improvements). Will they fix this bug?
> >
> > Upgrade and tell us.
> >
> > We don't put out bugfix releases just for the sake of wasting bandwidth.
> > If you're not running the latest version from your stable branch, IMHO
> > your bug reports should be ignored.
> >
>
> On the bug reporting form it says:
>
> Make sure you are running the latest available minor release for your
> major version before reporting a bug. The current list of supported
> versions is 8.2.5, 8.1.10, 8.0.14, 7.4.18.

Should we require the form to tell us their exact version number and
throw an error if it isn't the current one?

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://postgres.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +