logging and cron jobs

InfiniteLoop has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: logging and cron jobs by philcrow (Priest) on Oct 03, 2005 at 14:51 UTC
We use the syslog on Linux. It takes care of opening and closing the logs, ensuring that data is not lost. The Perl module is Sys::Syslog. I doubt this will work on non-unix systems. Phil	[reply]
Re: logging and cron jobs by adrianh (Chancellor) on Oct 03, 2005 at 16:49 UTC
As I understand, the log4perl will open a file handle to a log file and write into it. Some of the scripts, are run frequently and there might be cases of same script running as more than one process. In case of multiple processes, I fear the log file will be corrupted. Are my fears unfounded ? If you use Log::Log4perl::Appender::Synchronized then you shouldn't have any problems - it solves exactly this problem.	[reply]
Re: logging and cron jobs by Moron (Curate) on Oct 03, 2005 at 15:10 UTC
It all depends on what a neater solution means. I would try to understand the functional complaint before seeking technical alternatives. In this case, the issue is likely to be the lack of modularity of your solution, in the sense that you should be writing a logging subroutine that meets your manager's needs. Try to complete the list of requirements first, such as whether it needs a logfile naming convention (also creating required separation) to include timestamping, process identity, stage identity, file identity or whatever is functionally required. Design your own logging subroutine accordingly and get your manager to agree to it before coding it - you might find there are also site-specific requirements for how the logging routine is installed in a (site-specific) module as well as other project standards that need to be maintained. -M Free your mind	[reply]
Re: logging and cron jobs by Catharsis (Initiate) on Oct 03, 2005 at 16:19 UTC
in Log4perl you can do one log file per script by using a perl hook.. So in your log4perl config do.. log4perl.logger = DEBUG,LOGFILE log4perl.appender.LOGFILE = Log::Log4perl::Appender::File log4perl.appender.LOGFILE.filename = sub { return '/tmp/log.'.$$.'.log' } log4perl.appender.LOGFILE.layout = Log::Log4perl::Layout::PatternLayout log4perl.appender.LOGFILE.layout.ConversionPattern =%d %c - %m%n And to print out WHERE thats logging to do.. print Log::Log4perl->appender_by_name('LOGFILE')->filename();	[reply]
Re: logging and cron jobs by adrianh (Chancellor) on Oct 03, 2005 at 16:55 UTC
Also I would like your suggestion on other logger module(s). If you're just logging to files TheDamian's Log::StdLog is nice solution.	[reply]
Re: logging and cron jobs by idsfa (Vicar) on Oct 03, 2005 at 15:15 UTC
Ask your manager to define "neater". Separate activity logs for each run is very tidy and well organized. Shoehorning multiple, concurrent runs into a single log file is hard to follow even if you don't have corruption issues. Maybe you just need to have: `* 0 * * * cat /tmplogdir/date/temp.log >> /logdir/activity.log` [download] In the crontab as well The intelligent reader will judge for himself. Without examining the facts fully and fairly, there is no way of knowing whether vox populi is really vox dei, or merely vox asinorum. -- Cyrus H. Gordon*	[reply] [d/l]
Re: logging and cron jobs by sgifford (Prior) on Oct 03, 2005 at 15:59 UTC
One way to handle this is to create a seperate log file for every run of the program, using the time and PID to create a filename. To look through some or all of the logfiles, you can then use the shell's wildcards. For example, if you created logfiles named `yourapp.yyyy-mm-dd-hh-mm-ss.pid`, you could look at all the logs from 2005 with something like `cat yourapp.2005-*`. You could do this with a small wrapper program to manage the logfiles and run your program with `STDERR` connected to a pipe the wrapper program reads; that would require no changes to your current code.	[reply] [d/l]
Re: logging and cron jobs by kprasanna_79 (Hermit) on Oct 03, 2005 at 16:50 UTC
Hi infiniteloop, If you want to avoid STDERR word when ever you want redirect error to error file, you can tie a filehandle to a module where you can define print function. `untie STDOUT; $log = tie(STDOUT, 'module_name', "logfile_name") Inside the module sub print { my ($class, @data) = @_; print $class->{handle} = @data; . . .` [download] If you want to redirect both STDOUT and STDERR to file, please redirect these to a file.You can set the log file in runtime also. `open STDOUT, ">>$logfile"; open STDERR, ">&STDOUT";` [download] Tie this STDOUT to the module where you have implemented this print function. I think this resloves your issues. -Prasanna.K	[reply] [d/l] [select]