Test farm news

[Eric S. Raymond]

I now have a management script that can take a collection of ntp.conf files,
install them on farm machines, and start statistics gathering.  Real data has
been observed gathering in the logfiles.  The script can also collect all
logfiles for post-analysis.

That is, my first test launch looked like

farmer -s test1 fe.conf nd.conf ne.conf

and getting the data back looks like this:

farmer -c test1

The stats files will land in subdirectories named for the test hosts.
The ntpd log file is also copied.


Less happily, the Odroid people are failing to get their act together. They
tweaked their kernel build so gps-ppio is theoretically available, but our
own Dan Drown found that the support doesn't work.

He's sent them a patch, and I've sent them a rocket that if I don't get s
straight answer about how to query and set the GPIO pin binding *that I can
explain to newbies*, we're going to drop the Odroid.  Pointing me at devicetree
docs with an airy handwave usn't good enough, and I'm running out of patience.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>