Ports system quality
Alexander Leidinger
Alexander at Leidinger.net
Fri Sep 2 21:03:47 UTC 2011
Quoting Oliver Fromme <olli at lurza.secnetix.de> (from Fri, 2 Sep 2011
11:24:16 +0200 (CEST)):
> The other extreme are people who run a cron job every night
> that updates /usr/ports (*) and runs "400.status-pkg" (from
> /etc/periodic/weekly), possibly even followed by an automated
> update (**). Of course this will sometimes break. That's
> normal and to be expected, because the ports collection is
> changed and modified constantly by many people, except during
> freeze. There is always something that's broken. If you're
> affected, you need to postpone the update of the respective
> ports until someone (possibly including yourself) unbreaks it.
> That's the price to pay when you want to be on the "bleeding
> edge" instead of waiting for the next freeze and updating the
> ports to the release tag only.
No, that's the price to pay if you do not use all available tools.
Personally I make a FS snapshot before updating. We should recomment
to do this in all sensible places, and maybe even add code to
portmaster/portupgrade which tells to make a snapshot if there is none
(where possible). This way a rollback to a known good state is
possible if someone gets hit by an instability (someone still can get
hit, but the impact is a lot lower). Yes, I know that ZFS (the FS
where it is very easy to snapshot and rollback) is not used
everywhere, but the new installer for 9.0 offers now the possibility
and we should tell the people what is possible now.
Bye,
Alexander.
--
Important letters which contain no errors will develop errors in the
mail. Corresponding errors will show up in the duplicate while the
Boss is reading it.
http://www.Leidinger.net Alexander @ Leidinger.net: PGP ID = B0063FE7
http://www.FreeBSD.org netchild @ FreeBSD.org : PGP ID = 72077137
More information about the freebsd-ports
mailing list