Prev: "pg_stat_tmp/pgstat.stat": Stale NFS file handle
Next: [HACKERS] Nice hint on table aliasing!
From: Tom Lane on 4 Apr 2010 12:19
Hitoshi Harada <umi.tanuki(a)gmail.com> writes:
> I cannot figure out at all what is wrong. Have any idea?

Since nobody else is reporting this, it seems like it must be either
something messed up about your system, or something wrong with your
copy of the PG sources. In the latter connection I confess to having
little confidence in the git mirror setup. Have you tried diffing
your source tree against a CVS checkout or nightly snapshot tarball?

A different line of attack is to insert some debugging logging into
the postmaster's port-opening code to see if you can find why it's
apparently not doing anything.

regards, tom lane

--
Sent via pgsql-hackers mailing list (pgsql-hackers(a)postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

From: Andrew Dunstan on 4 Apr 2010 12:52


Hitoshi Harada wrote:
> The problem isn't in libpq, since it is that the server doesn't listen
> on startup as above. No /tmp/.s.PGSQL.5432, nor LISTENING entry in
> netstat.


I find this somewhat implausible. The postmaster has this code that
makes it die if it can't open a listening socket:

if (ListenSocket[0] == PGINVALID_SOCKET)
ereport(FATAL,
(errmsg("no socket created for listening")));

Perhaps you could strace the execution.

Of course, pg_regress doesn't usually run on port 5432, IIRC, so maybe
you're looking for the wrong thing.

Another question worth asking is whether or not either SELinux or
firewall settings on your CentOS box are interfering with connectivity.

cheers

andrew





--
Sent via pgsql-hackers mailing list (pgsql-hackers(a)postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

From: Josh Berkus on 4 Apr 2010 13:13
On 4/4/10 9:19 AM, Tom Lane wrote:
> Hitoshi Harada <umi.tanuki(a)gmail.com> writes:
>> I cannot figure out at all what is wrong. Have any idea?

We ran Make Check on 4 linux laptops and 3 macs yesterday, and did not
see a hang.

Could this be an issue with VirtualBox? Have you used this VM for
testing before?

--
-- Josh Berkus
PostgreSQL Experts Inc.
http://www.pgexperts.com

--
Sent via pgsql-hackers mailing list (pgsql-hackers(a)postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

From: Hitoshi Harada on 4 Apr 2010 20:54
2010/4/5 Tom Lane <tgl(a)sss.pgh.pa.us>:
> Hitoshi Harada <umi.tanuki(a)gmail.com> writes:
>> I cannot figure out at all what is wrong. Have any idea?
>
> Since nobody else is reporting this, it seems like it must be either
> something messed up about your system, or something wrong with your
> copy of the PG sources.  In the latter connection I confess to having
> little confidence in the git mirror setup.  Have you tried diffing
> your source tree against a CVS checkout or nightly snapshot tarball?

I've tried clean tarball of alpha4, alpha5, so source messing seems
less possible. I'll try CVS HEAD. I've check various real (not VM)
Linux boxes and didn't hit the same situation. I know it's something
around environment issue.

>
> A different line of attack is to insert some debugging logging into
> the postmaster's port-opening code to see if you can find why it's
> apparently not doing anything.

I'll try it later. But debugging with gdb seems quite clean, bind()
and other initializations looks good and postmaster is in
ServerLoop().

Regards,


--
Hitoshi Harada

--
Sent via pgsql-hackers mailing list (pgsql-hackers(a)postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

From: Hitoshi Harada on 4 Apr 2010 21:28
2010/4/5 Josh Berkus <josh(a)agliodbs.com>:
> On 4/4/10 9:19 AM, Tom Lane wrote:
>> Hitoshi Harada <umi.tanuki(a)gmail.com> writes:
>>> I cannot figure out at all what is wrong. Have any idea?
>
> We ran Make Check on 4 linux laptops and 3 macs yesterday, and did not
> see a hang.
>
> Could this be an issue with VirtualBox?  Have you used this VM for
> testing before?
>

Yeah, this VM has mostly been used to develop my window function's
patches, so this is critical for me.

Regards,

--
Hitoshi Harada

--
Sent via pgsql-hackers mailing list (pgsql-hackers(a)postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

 |  Next  |  Last
Pages: 1 2
Prev: "pg_stat_tmp/pgstat.stat": Stale NFS file handle
Next: [HACKERS] Nice hint on table aliasing!