Re: [Savannah-hackers-public] [gnu.org #374779] Commits to the "www" web repository do not propagate

[Sylvain Beucler via RT]

On Fri, Sep 05, 2008 at 11:17:23AM -0400, Ward Vandewege via RT wrote:
> > [yavor - Wed Sep 03 19:21:03 2008]:
> > 
> > Joakim Olsson wrote:
> > > 
> > > but nothing happened at http://www-test.gnu.org/fun/jokes/gnuemacs.html
> > 
> > Since yesterday at around 15:00-17:00 UTC (based on non-scientific
> > observation) commits in the "www" project do not appear online at
> > nadesico.
> > 
> > This happened many times before, and was sysadmin's concern -- but now
> > I don't know exactly how it's handled since the replication mechanism
> > changed recently.
> > 
> > I'm CC-ing both the GNU sysadmins and the Savannah hackers for extra
> > safety, although I anticipate the former should undertake action.
> 
> I temporarily disabled the cvs update cron jobs because of the load that
> the Fry video launch generated. They have been enabled again since
> sometime yesterday.
> 
> This was also a test of the 'push' updates that should now happen from
> Savannah whenever there is a checkin for the www project. 
> 
> Apparently, those don't work yet. Savannah hackers, can you look into
> this? I'd like to retire the cvs update cron jobs permanently.

Hmmm,
> > > but nothing happened at
> > > http://_www-test_.gnu.org/fun/jokes/gnuemacs.html

'www-test' is disabled, now that we have proper on-commit replication.

Yavor, are you talking about 'www-test' or 'www' (or both)?


For reference here's the CVSROOT/loginfo:
  DEFAULT /usr/local/bin/log_accum.pl --mail-to address@hidden --send-diff 
--file-prefix www_106 --config /etc/log_accum-web.config %p %{sVv}
  ALL echo 'Triggering webpages update...'; cat > /dev/null; curl 
http://www.gnu.org/new-savannah-project/new.py -s -F type=www -F 
project=`basename %r`

So basically all commits that appear on address@hidden also
triggered a replication.


I'd like, too, the replication cron to be disabled - I thought this
was already done.  Can you do that now, so that we can really test
whether the system works?  Let us know when you do.

-- 
Sylvain