Apache and Logrotate

[Brian Curtis <bcurtis pomfret net>]

Hello,

I've been having intermittent problems with Apache and Logrotate.
I need to rotate over 100 logfiles on a weekly basis and about another
dozen on a daily basis.

It seemed to me that -HUP'ing the httpd process after every logfile
rotation was a bit excessive, and sometimes caused problems when the
process did not completely stop before the next -HUP.

I've tried doing a stop then restart on the last logfile rotation
entry with a 15 second pause in between to make sure the process exits
before trying to restart.  This usually works, but it has the same
drawbacks as -HUP'ing the process after every log rotation -- if it
fails, there's problems.

Basically I'm looking for comments and suggestions from those of you
who might have either experienced the same thing, and/or already found
a better way to deal with numerous log rotations on a weekly and daily
basis.


Also, how are you handling worm-infected server probes?  We're still
seeing quite a bit of Nimda, Code Red, etc. infected servers making
several requests per second probing for vulnerabilities.

While not major problem, it does waste quite a bit of bandwidth and
is quite annoying since these IIS "admins" have had a good deal of
time to patch their servers.

TIA for all suggestions.

-- 
Best regards,
 Brian Curtis