Re: beta2 rpms

From: Lamar Owen <lowen(at)pari(dot)edu>
To: Jon Jensen <jon(at)endpoint(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: beta2 rpms
Date: 2004-09-02 21:08:16
Message-ID: 200409021708.16258.lowen@pari.edu
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Thursday 02 September 2004 16:27, Jon Jensen wrote:
> I've been meaning to ask for a long time: Why does /var/log/pgsql get
> installed with the execute bit set? I don't have any other log files with
> the execute bit on, and can't imagine why that would be necessary or
> useful. Am I missing something?

Hmmmm.... I've set the wrong permission mask in the %files list, apparently.

> Also, is there a reason that the init script defaults to redirecting
> stdout and stderr to /dev/null instead of to /var/log/pgsql?

The intent was t use syslog. The new log rotation scheme will likely be
tapped this time around, though.

> If the DBA
> doesn't turn on logging in postgresql.conf, then very little output goes
> to /var/log/pgsql, usually of an important nature like "stopped" or
> "started" at a certain date. I normally change that first thing after a
> new install.

Normally the first thing I do is enable and set up syslog support. This way I
can use my established syslog infrastructure, where really important log
messages not only get logged to disk but get PRINTED. Logs that print are
difficult for intruders to mess with, and once one has an established syslog
server it's trivial to send all logs from all servers to the one central
logging server. I have done this a long time.
--
Lamar Owen
Director of Information Technology
Pisgah Astronomical Research Institute
1 PARI Drive
Rosman, NC 28772
(828)862-5554
www.pari.edu

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Dann Corbit 2004-09-02 21:13:36 Re: Implementation of MMDBMS
Previous Message Gaetano Mendola 2004-09-02 20:47:16 Re: version upgrade