Re: pg_group_name_index corrupt? - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: pg_group_name_index corrupt?
Date
Msg-id Pine.LNX.4.21.0005071301001.442-100000@localhost.localdomain
Whole thread Raw
In response to Re: pg_group_name_index corrupt?  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: pg_group_name_index corrupt?  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane writes:

> > BTW, currently the best way to get logs from postmaster woul be
> > compiling it with USE_SYSLOG (why is this not enabled by configure?),

An --enable-syslog switch will be in the next release unless someone
protests.

> Good question.  Before we standardize on that, however, some testing
> might be in order.  I haven't tried stuffing multimegabyte querytree
> dumps into syslog ... will it work?  On how many platforms?  The
> syslog applications I've seen never write more than a hundred or so
> characters per log entry, so I'm a tad nervous about assuming that we
> will get reasonable behavior for large messages...

Indeed. My syslog becomes a little erratic with messages longer that half
a screen or so. Nothing critical but it splits up the message in separate
chunks together with `Cannot glue message parts together' entries. It also
sends out an empty log message to everyone's terminal. I recall that while
doing some work for getting exim acquainted with syslog we have had the
same problems with long messages. Also, I bet this syslog implementation
(claims to be derived from "stock BSD") is very widespread.

It's debatable whether these sort of things are appropriate for syslog,
though. Who cares about the parse tree the next day?

-- 
Peter Eisentraut                  Sernanders väg 10:115
peter_e@gmx.net                   75262 Uppsala
http://yi.org/peter-e/            Sweden



pgsql-hackers by date:

Previous
From: Peter Eisentraut
Date:
Subject: Re: CREATE DATABASE WITH OWNER '??';
Next
From: Peter Eisentraut
Date:
Subject: Re: You're on SecurityFocus.com for the cleartext passwords.