httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Plüm, Rüdiger, VF-Group" <ruediger.pl...@vodafone.com>
Subject Re: svn commit: r733493 - in /httpd/httpd/trunk: CHANGESdocs/man/rotatelogs.8 docs/manual/programs/rotatelogs.xmlsupport/rotatelogs.c
Date Fri, 16 Jan 2009 11:08:29 GMT
 

> -----Ursprüngliche Nachricht-----
> Von: Rainer Jung 
> Gesendet: Freitag, 16. Januar 2009 11:51
> An: dev@httpd.apache.org
> Betreff: Re: svn commit: r733493 - in /httpd/httpd/trunk: 
> CHANGESdocs/man/rotatelogs.8 
> docs/manual/programs/rotatelogs.xmlsupport/rotatelogs.c
> 
> On 12.01.2009 10:04, Joe Orton wrote:
> > On Sun, Jan 11, 2009 at 05:36:07PM -0000, rjung@apache.org wrote:
> >> Author: rjung
> >> Date: Sun Jan 11 09:36:07 2009
> >> New Revision: 733493
> >>
> >> URL: http://svn.apache.org/viewvc?rev=733493&view=rev
> >> Log:
> >> Allow to trigger rotatelogs log file rotation from
> >> using HUP and INT signals to the rotatelogs process.
> >>
> >> This is helpful, when log activity is low, but you want
> >> rotatelogs to close the open log files.
> >
> > Sending SIGTERM to the rotatelogs process and having the 
> parent recycle
> > it should have done that already, surely, without adding all this
> > complexity to rotatelogs?
> >
> > A lot of the stuff you're doing in the new signal handler is not
> > async-signal safe, so -1.  (closeFile notably calls pool functions)
> 
> I send a patch for trunk to the list, including reliable 
> piped logging 
> for the error loggers.
> 
> Coming back to "adding complexity": as you might have noticed 
> I added a 
> bit of other functionality to rotatelogs and part of the complexity 
> comes from making it a bit more modular (moving code into functions).
> 
> The added functionality is:
> 
> a) allow to rotate based on time *and* size (check both and rotate 
> whenever one of the two criteria applies)
> 
> b) verbose output (for debugging purposes, log parsed config 
> flags and 
> whenever rotation occurs)
> 
> c) Allow B, K, G, T as unit characters for sizes in addition to 'M'
> 
> I really like a) but b) and c) are not that important. 
> Removing b) makes 
> the code shrink about 60 lines.
> 
> Any preferences?

As far as I understand Joe he is mostly concerned about the signal code.
So I guess reverting r733493 should be sufficient. All other stuff should
be ok.

Regards

Rüdiger

Mime
View raw message