Date: Fri, 30 Jan 2004 10:53:41 -0500 From: Ken Smith <kensmith@cse.Buffalo.EDU> To: John Baldwin <jhb@freebsd.org> Cc: Daniel Eischen <eischen@vigrid.com> Subject: Re: cvs commit: src/lib/libc_r Makefile src/lib/libpthread Makefile (fwd) Message-ID: <20040130155341.GA13720@electra.cse.Buffalo.EDU> In-Reply-To: <200401301021.08373.jhb@FreeBSD.org> References: <Pine.GSO.4.10.10401300716370.7499-100000@pcnet5.pcnet.com> <200401301021.08373.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Jan 30, 2004 at 10:21:08AM -0500, John Baldwin wrote:
> On Friday 30 January 2004 07:22 am, Daniel Eischen wrote:
> > I built world on alpha to test this, but I couldn't install it
> > to make sure the links from libpthread.* -> libc_r.* were made
> > correctly. Could someone please verify this? Feel free to
> > fix it if it is wrong (src/lib/libc_r/Makefile).
>
> You can do an installworld into a DESTDIR=/junk to make sure the links are set
> correctly.
Only if you have root on the box, things like:
install -o root -g wheel <blah>
start failing pretty quick if you don't.
If you're in that boat a "worst case" approach to this would be
checking the cluster reference machines to see what they thought
of the change the next day. Peter set up cron jobs on each of the
reference machines in the wee hours of the morning that do a
cvs update of /usr/src, "make buildworld", then rebuild the
kernel, install kernel/world, and reboot. If any of those steps
fail the script exits without the reboot.
So, use ruptime to see how long the machine has been up to make
sure the rebuild didn't fail, and if it's been up less than 24
hours check to see if what you wanted to happen did.
beast is the Alpha ref machine, panther is the sparc64 ref machine,
and sledge is the amd64 machine. At the moment the two ia64 machines
do NOT do this, but marcel keeps them pretty close to -current all
the time. The i386 ref machine for -current is ref5 but it's down
due to hardware issues at the moment.
I usually don't worry too much about a ref machine skipping a day,
if any of them have been up more than 2 days I usually poke through
the log files to see why, and perhaps nag someone if it seems they
might be ignoring tinderbox reports. :-) If any of them are crashing
due to the newest kernel I deal with booting them on the previous
day's kernel (and usually save a copy of that in case nobody fixes
the problem before the next run of the cron job :-).
--
Ken Smith
- From there to here, from here to | kensmith@cse.buffalo.edu
there, funny things are everywhere. |
- Theodore Geisel |
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040130155341.GA13720>