Wow, last time I've heard about JKL Toys was ages ago on Ignite/400!
If I'm not wrong, Log cicles end and starts at midnight, so you can see some
HTTP jobs ending and starting that time. Some eventually also ends and
starts on other times, but log is not affected.
Don't know if this time (midnight) can be changed (LogMaintHour?) and don't
know how to configure those jobs restarts (not on midnight).
Years ago I've asked support on how to change log file names (ex:Q11408200)
the answer was I couldn't. It seens that log rotation is more a question of
OS than Apache, so guys at Rochester developed a module that doesn't work
exactly as on other platforms. But that was ate JKL time, maybe now is
When Apache is running (on any OS) current log file is always open, so
saving it causes problems, but there's no problem with older log files, they
can be saved without issues.
As they are very big, we don't maintain log files on the server, so we
transfer them to a PC. Yesterday's closed log is done today and so on, no
logs are saved on the iSeries side.
LogMaint directives are responsible for logs cleaning and control, but it
seems that some Apache functions are not 100% implemented on iSeries, so
test them before applying on any production environment.
We have a couple of HTTP instances only to test variations and
configurations, most based on Apache's own documentation and other forums or
lists, there's a lot of material and almost all works fine or there's
another way to work fine, as usually different directives can be combined to
reach about same results.
Some time ago we've done some testing on Apache based on Windows, found
similar little issues on other features... Seems nobody's perfect!
De: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] Em nome de
Enviada em: quarta-feira, 20 de agosto de 2014 16:42
Assunto: HTTP instance issues / config questions / documentation
1) I see a HTTP instance job always starting/ending every night at midnight
for most instances, but not all. What directive is controlling this? What is
the purpose of this job?
2) Some of my HTTP instances "auto clean" their log files, others don't.
Are these the directives responsible for this?
LogMaint logs/access_log 30 0
LogMaint logs/error_log 30 0
3) Within the HTTP instance logging job, I'm occasionally receiving the
following error messages:
a) HTP8424-Error occurred opening log file
b) HTP8431 - Logging job for HTTP server RXSTIVO failed.
c) CPF24A3 - Value for call stack counter parameter not valid.
This occurs occasionally at midnight, while my daily BRMS save is saving the
IFS. Timing issue - This is because the save has a brief lock on the IFS WWW
folder. Once this occurs, logging job fails, no longing for these
Can only the logging job be restarted, or does it take a total recycle of
the HTTP instance?
I'm reviewing various options from BRMS group on how to correct this.
Also waiting for confirmation on IFS locking rules during a save.
One option I recently discovered is to move the HTTP maint window away from
midnight using the LogMaintHour directive.
4) All of our HTTP instances were pre-installed by 3rd party apps, never
really reviewed for current admin standards/updates/security/logging, etc.
Besides the Apachedft instance or the example from JKL Toy Company, are
there any good examples of a typical HTTP instance?
Actually, some of our HTTP instance almost mirror the JKL Toy Company
example (which is now quite old).
I'd like to bring the HTTP configs up to some sort of standard, but hesitant
because something may break/fail.
5) I've located 3 IBM I HTTP documents.
Are there any other good recent docs?
IBM i: e-business and Web Serving IBM HTTP Server for I pdf
IBM i Systems Administrator
Pencor Services, Inc.
462 Delaware Ave
Palmerton Pa 18071