[check_postgres] [commit] Two cool new checks: new_version_cp and new_version_pg to check if you

[Robert Treat]

On Wednesday 08 April 2009 04:12:54 Cédric Villemain wrote:
> Guillaume Lelarge a écrit :
> > Le mercredi 08 avril 2009 à 09:44:04, Cédric Villemain a écrit :
> >> check_postgres at bucardo.org a écrit :
> >>> Committed by Greg Sabino Mullane <greg at endpoint.com>
> >>>
> >>> Two cool new checks: new_version_cp and new_version_pg to check if you
> >>> have the latest and greatest check_postgres.pl and that you are running
> >>> the latest revision of Postgres.
> >>
> >> Very nice !
> >>
> >> I just wonder why use the docs/ instead of the ftp/source ? Perhaps to
> >> check for very o(l)dd version of postgresql ?
> >>
> >> url : http://www.postgresql.org/ftp/source/
> >> look for greatest <a href>: /ftp/source/v(8\.3.*)/
> >>
> >> It is then possible to issue a CRITICAL if the current server is not
> >> maintain anymore by PGDG (pg < 7.4).
> >
> > Using the RSS feed could be a better way:
> >
> >   http://www.postgresql.org/versions.rss
>

+1. The rss feed is the web teams "official" tool for this type of thing, so 
that is really what we should be using. 

> Per offline talk with Guillaume :
>
> The ftp is first updated then the feed. And the feed is used for the main
> page of the website (so ftp can be updated hours/day before the official
> announce, and we want to follow official announce, right?)
>

Yes, especially given that we (postgres packaging team) occasionally upload 
packages into the ftp site that get revoked / pulled down before official 
release announcement because of problems with the package/code. Sometimes 
this results in a re-wrap and version bump, but not always, which means if 
you are grabbing from the ftp site before the announcement goes out, you 
might grab an invalid release; I don't think we want to encourage that. 

BTW, yes, those last sentences are scary. Don't shoot the messenger. 

> Some interesting point with the feed is the ability to catch the date of
> the lastest release. It can let us catch the delay since the release and
> make Warning and Critical adjustable :
>
> Critical if lastest release older than X days for example.
>

Hmm... this gives me an idea... possibly we (webteam) could modify the rss 
feed to include whether a release includes security fixes, so that we 
(check_postgres) could adjust nagios check based on that information. 

-- 
Robert Treat
Conjecture: http://www.xzilla.net
Consulting: http://www.omniti.com