Re: [logcheck] I hear you...
On Sun, Nov 25, 2001 at 02:58:04PM -0800, Karl M. Hegbloom wrote:
> Darren> And the scripts or binaries called from them to write
> Darren> errors on stderr. I note that apachectl (apache 1.3.22-2)
> Darren> writes errors to stdout - hmm, bug report?
>
> Let's go read LSB, and decide. You write the report if appropriate, Ok?
What he's basically saying is that everything suggested here is stated
in the LSB, so you can go ahead and make a fancy-sysinit program and if
something doesn't work as expected, you can blame it on LSB non-compliant
init scripts (or is that non-LSB compliant).
I havn't looked over that code you sent too much but it doesn't seem to
handle interleaved stdout and stderr. I think someone should come up with a
really cool design to make everyone happy. You could have a
fancy-sysvinit.conf looking like:
bootstyle = classic | redhat | hpux | ...
colo[u]r = yes | no
okfailnote = before | after | flushright | nextline | ...
stdout = white | lightgrey | ...
stderr = yellow | blue | ...
errormailto = root@localhost
I think that should be able to make everyone happy. Ofcourse, you want to be
able to do it no matter what boot method people use (sysvinit, filerc, etc)
which would mean a separate program (exec-init.d?? service??) that they can
all use.
Unfortunatly, it still seems that trapping kernel messages before klogd has
started could be tricky...
So, anyone voluteering? :)
--
Martijn van Oosterhout <kleptog@svana.org>
http://svana.org/kleptog/
> Terrorists can only take my life. Only my government can take my freedom.
Reply to: