[collectd] 4.2.1 on Solaris 10

[Christophe Kalt]

On Dec 05, Sebastian Harl wrote:
| _FILE_OFFSET_BITS gets defined by the AC_SYS_LARGEFILE autoconf macro, thus
| I'd consider this to be a bug in autoconf. As a temporary workaround we could
| disable the macro for Solaris - that'll be a bad hack though, which I don't
| really like.
|
| You should file a bug against autoconf (I assume you're using GNU autoconf).

I'm using the configure script that is shipped by collectd.
Granted, i didn't bother to look at this in detail as
disabling that particular plugin is quick & easy.  But other
things compile just fine, so saying that it's an autoconf bug
may be a bit premature.

| D'oh... a lot of those *_r functions seem to be quite system-dependent. Can
| you please provide the definition of getgrnam_r() as defined on Solaris
| (simply providing the appropriate part of the manpage should be fine)?

http://docs.sun.com/app/docs/doc/816-5168/getpwnam-r-3c?a=view

| As mentioned above LFS is enabled by using the AC_SYS_LARGEFILE autoconf
| macro.

Sure, and that's the right way to do so.

|        Can you please check how "swap -s" gets the values on Solaris? Forking
| a process each interval is a rather bad idea...

Easy, it calls swapctl(2); but unlike collectd, swap(1M) does
not also need LFS support.  I agree that forking constantly
is not a good idea, but the only choice you have here (as far
as i can tell) is to either fork each interval, or have a
separate daemon process.
My quick read of the collectd 4 docs seem to indicate the
latter can easily be achieved with the exec and unixsock
plugins, but what i was saying is that it'd be nicer if
collectd did that transparently.

Christophe