Hi Alex, On 6/4/07, Alexander Clouter <ac56@soas.ac.uk> wrote:
Hi,
Luiz Cruz <luizcruz.msg@gmail.com> [20070604 09:43:08 -0200]:
Hi all,
I have a doubt about a process that i have. Maybe the solution is in the side of syslog-ng, i am searching. My process is:
Log -> Perl Script -> Fifo -> Syslog-ng -------------- > Syslog-ng -> perl -> rrd
Can't this be redone as:
Perl Script ---(Sys::Syslog)---> syslog-ng -> syslog-ng -> Perl -> RRD
You do not explain where and what 'Log' is so it is difficult to help out.
Nice, don´t use a fifo file sounds great. I will try. Log is from a jetty aplication. The request is proxyed by apache. But i have to parse the lines before sent to rrd (cacti). The log is about 1GB/day. Sorry if i didn´t mention before.
The problem is when the application write to many lines in the fifo file (through the perl script), my syslog-ng (Standard Configuration) can't process all lines quickly, and the load of cpu rises a lot. Is there any parameters that i can use to optimize the reading? Someone have a similar solution?
If 'Log' is a file on the machine why can you not just use the 'tail' functionality[1] of syslog-ng?
Cheers
Alex
[1] http://www.balabit.com/dl/html/syslog-ng-admin-guide_en.html/ch08s01.html#id...
I didn´t find this... is that a pipe function?
Thanks a lot.
Att., Luiz
Again. Thanx! Luiz