handler for error log

[John Saylor <jo...@worldwinner.com>]

Hi

Is it possible to put a handler on the error log so that certain
elements could be filtered? Ideally, I'd like to keep track of how many
times a certain error appears and with a handler that would be a cake
walk. So the handler would 'live' between the main httpd process and the
file, doing it's magic there.

I've done some searching and found nothing [other than people
complaining about their CGI scripts].

Any pointers or suggestions are welcome.

-- 
\js

Wayne's World C Programming Style Guide:
	A == B; !;

("A equals B. Not!")

Re: handler for error log

[Steven Lembark <le...@wrkhors.com>]

> Is it possible to put a handler on the error log so that certain
> elements could be filtered? Ideally, I'd like to keep track of how
> many times a certain error appears and with a handler that would be a
> cake walk. So the handler would 'live' between the main httpd process
> and the file, doing it's magic there.
>
> I've done some searching and found nothing [other than people
> complaining about their CGI scripts].
>
> Any pointers or suggestions are welcome.

There are several log watchers avaialble, including ones written
in perl.  One simple trick is to essentially tail -f the thing into a
regex.  Quick way is to seek to the end, store -M $logfile and set
$\ = undef.  After that:

	open my $logfile, "< $logpath" or croak "$logpath: $!";

	undef $\; # or leave it alone if reading into an array.

	for( my $lastmod = -M $logfile ;;)
	{
		sleep 10 while( $lastmod == -M $logfile );

		my $newstuff  = <$logfile>;

		# regex $newsuff to your heart's delight.
		# that or read it as @newstuff and process it
		# line by line
	}


		



--
Steven Lembark 
2930 W. Palmer
 
chicago, IL 60647
lembark@wrkhors.com 
+1 800 762 1582

Re: handler for error log

[Doug MacEachern <do...@covalent.net>]

On Tue, 12 Jun 2001, John Saylor wrote:

> Hi
> 
> Is it possible to put a handler on the error log so that certain
> elements could be filtered? Ideally, I'd like to keep track of how many
> times a certain error appears and with a handler that would be a cake
> walk. So the handler would 'live' between the main httpd process and the
> file, doing it's magic there.

you can use a piped log to feed error_log output to a script.  there are
no in-process error_log hooks for 1.x

see: http://httpd.apache.org/docs/mod/core.html#errorlog
"If the filename begins with a pipe (|) then it is assumed to be a command
to spawn to handle the error log."