----- Original Message -----
> > [...]
> > So, we are talking about "should we use pmlogconf or cat-a-series-of-all-
> > possible-configuration-files, with no knowledge of the remote host setup,
> > and let pmlogger sort it out when it starts up" to manage our farm of
> > pmlogger instances [...]
>
> Not necessarily. pmlogconf could still be run for each newly discovered
> remote host; its output file could be in a separate per-host config file.
That's how the pmlogconf and pmie_check scripts integration works already.
I think we're talking past each other a bit here - you may be seeing a
problem which I think doesn't exist. :)
> By the way, another (non-pmcpp, non-directory-searching) simple way to
> mix hand-written and generated files would be to permit pmlogger to
> have not just one -c CONFIG file option, but multiple.
Yep, indeed. As with the cat-all-in-a-directory option though, my first
thoughts are that it might not add a lot of value, for the same reasons -
pmlogconf solved this problem a few years ago and in a more flexible way
(IMO).
Reflecting further, one area where I wonder if we would get alot of extra
value would be to add cat-a-series-of-possible-configs-into-one support
into the /etc/pcp/pmlogger/control file. As new hosts come and go, if
we could add/remove files in a subdirectory rather than having to edit a
single file, I think it would make the Avahi event handling process(es)
far simpler. Shouldn't be too hard either - that file is versioned, so
we could bump the version and add an "include" keyword... then, perhaps
/etc/pcp/{pmie,pmlogger}/control.d directories for the Avahi responder(s)
to write into?
cheers.
--
Nathan
|