Обсуждение: logging problem ... ?
Okay, I'm at a loss here ... I'm trying to futz with pg_options, based on
the docs in users-lounge/7.0/docs, and with:
pgsql% cat data/pg_options
verbose=2
query
hostlookup
showportnumber
All I'm getting in my log file is this continuously scrolling:
FindExec: found "/pgsql/bin/postgres" using argv[0]
FindExec: found "/pgsql/bin/postgres" using argv[0]
FindExec: found "/pgsql/bin/postgres" using argv[0]
FindExec: found "/pgsql/bin/postgres" using argv[0]
FindExec: found "/pgsql/bin/postgres" using argv[0]
even if I issue a query through psql, that is all I see ...
Is this broken? *raised eyebrow*
Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
The Hermit Hacker <scrappy@hub.org> writes:
> All I'm getting in my log file is this continuously scrolling:
> FindExec: found "/pgsql/bin/postgres" using argv[0]
> FindExec: found "/pgsql/bin/postgres" using argv[0]
Well, that shows you're launching backends anyway ;-)
It looks like the backends are receiving -d 1 from the postmaster,
but where is the postmaster's own debug output? Weird. How exactly
are you starting the postmaster, again?
regards, tom lane
On Mon, 8 May 2000, Tom Lane wrote:
> The Hermit Hacker <scrappy@hub.org> writes:
> > All I'm getting in my log file is this continuously scrolling:
>
> > FindExec: found "/pgsql/bin/postgres" using argv[0]
> > FindExec: found "/pgsql/bin/postgres" using argv[0]
>
> Well, that shows you're launching backends anyway ;-)
Ya, but, from what i can tell, there are not as many backends being
started as that scrolling list of 'FindExec's appear to be indicating
... the server isn't *that* busy ...
> It looks like the backends are receiving -d 1 from the postmaster,
> but where is the postmaster's own debug output? Weird. How exactly
> are you starting the postmaster, again?
cat ~pgsql/pgstart ... postmaster's own debug output? ya know, for some
stupid reason, the only thing i'm capturing is postgres's output :( I
just modified the pgstart script and restarted the server, so postmaster's
output is to logs/postmaster.5432 ...
Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
The Hermit Hacker <scrappy@hub.org> writes:
> cat ~pgsql/pgstart ... postmaster's own debug output? ya know, for some
> stupid reason, the only thing i'm capturing is postgres's output :( I
> just modified the pgstart script and restarted the server, so postmaster's
> output is to logs/postmaster.5432 ...
Oh, I see the problem! It's the -S in the postmaster switches:
${POSTMASTER} -d 1 -B 4096 -N 128 -S \ ^^
That redirects postmaster's stdout/stderr to /dev/null, so you never
see anything in its log except for a message or two that comes out
before -S is processed. Instead of -S you should be using
nohup ${POSTMASTER} same-switches-except-S \</dev/null >& logs/postmaster.${PORT} &
(I'm not sure if redirecting stdin to /dev/null is really necessary on
your OS, but it is on mine.)
regards, tom lane
D'oh ... there ya go, lotsa days flying through now ...
On Mon, 8 May 2000, Tom Lane wrote:
> The Hermit Hacker <scrappy@hub.org> writes:
> > cat ~pgsql/pgstart ... postmaster's own debug output? ya know, for some
> > stupid reason, the only thing i'm capturing is postgres's output :( I
> > just modified the pgstart script and restarted the server, so postmaster's
> > output is to logs/postmaster.5432 ...
>
> Oh, I see the problem! It's the -S in the postmaster switches:
>
> ${POSTMASTER} -d 1 -B 4096 -N 128 -S \
> ^^
>
> That redirects postmaster's stdout/stderr to /dev/null, so you never
> see anything in its log except for a message or two that comes out
> before -S is processed. Instead of -S you should be using
>
> nohup ${POSTMASTER} same-switches-except-S \
> </dev/null >& logs/postmaster.${PORT} &
>
> (I'm not sure if redirecting stdin to /dev/null is really necessary on
> your OS, but it is on mine.)
>
> regards, tom lane
>
Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
The Hermit Hacker <scrappy@hub.org> writes:
> D'oh ... there ya go, lotsa days flying through now ...
Ah, much better. You just *thought* you weren't launching a backend
every second or so :-)
Kinda looks like the backend -o switch doesn't redirect as much as
I would've thought --- most of the backend's debug output is coming
out in the postmaster log not the other file. Something to look at
later.
regards, tom lane
On Mon, 8 May 2000, Tom Lane wrote:
> The Hermit Hacker <scrappy@hub.org> writes:
> > D'oh ... there ya go, lotsa days flying through now ...
>
> Ah, much better. You just *thought* you weren't launching a backend
> every second or so :-)
tell me about it ... now, if it were possible to redirect each backend to
a seperate file, so that I had X files pertaining to what happened through
that backend, that would be cool ... maybe play with v7.1 on this one :)
> Kinda looks like the backend -o switch doesn't redirect as much as
> I would've thought --- most of the backend's debug output is coming
> out in the postmaster log not the other file. Something to look at
> later.
*nod*
Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
The Hermit Hacker writes: > tell me about it ... now, if it were possible to redirect each backend to > a seperate file, so that I had X files pertaining to what happened through > that backend, that would be cool ... maybe play with v7.1 on this one :) You can compile with ELOG_TIMESTAMPS to tag each log message with a pid (besides the timestamp). > > Kinda looks like the backend -o switch doesn't redirect as much as > > I would've thought --- most of the backend's debug output is coming > > out in the postmaster log not the other file. Something to look at > > later. This might explain it: peter ~/pgsql/src/backend$ find -name '*.c' | \ xargs egrep 'fprintf *\(stderr' | wc -l 232 -- Peter Eisentraut Sernanders väg 10:115 peter_e@gmx.net 75262 Uppsala http://yi.org/peter-e/ Sweden