Thread: Re: Industrial-Strength Logging

Re: Industrial-Strength Logging

From
Bruce Momjian
Date:
Can someone tell me what happened to this?


>
> This is NOT a patch file! Read the README files for info on how to
> apply the contents to a Dec 4 or later postgresql.snapshot.
>
> This a preview release and is neither clean nor completely portable, but
> it has all the basic functionality so you can see how it works and where
> it's headed.
>
> You SHOULD be able to create and run PostgreSQL with logging functionality
> under Linux (especially Red Hat V6. If not, PLEASE contact me!)
>
> You can PROBABLY get it going under any SysV-compatible Unix.
>
> BSD users probably only get to admire the source code for the moment.
>
> I'm available at mtsinc@southeast.net or on the pgsql-hackers newsgroup
> for questions, comments, and abuse.
>
> See http://216.199.14.27/ for docs.
>
> Thanks to those who made me feel like the effort would be appreciated and
> to those whose suggestions and comments helped guide the design and development.
>
>    Tim Holloway

[ application/x-gzip is not supported, skipping... ]


--
  Bruce Momjian                        |  http://www.op.net/~candle
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

Re: Industrial-Strength Logging

From
Peter Eisentraut
Date:
Bruce Momjian writes:

> Can someone tell me what happened to this?

Well, as far as I followed along we at first thought that using syslog for
logging would be fine. Then we found out that syslog is really quite
unusable. Now that is fixed, but only in a work-around kind of
fashion. Meanwhile someone came up with the idea of abducting rotatelogs
from Apache, which seems like an interesting thing to do in any case. At
the same time, it turned out that we already have a fair amount of logging
information available, you just need to turn it on.

The problems people saw with this approach in particular are the separate
logging process, the pretty deep surgery in the system in general, and the
use of non-standard error or event codes. It seems that this project went
from requirements definition to implementation without anyone knowing what
happened in between.

Yeah, let's have another logging discussion... :)


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