[lxc-devel] How close are we to 0.9.0 alpha3 or better?

[Stéphane Graber]

On 01/19/2013 06:53 PM, Michael H. Warfield wrote:
> Hey all...
> 
> I don't see that we've got a firm roadmap in place of milestones to
> cross or blockers to get out of our way but...
> 
> I had this concern when we cut 0.8.0, without certain critical systemd
> features, that the next cut would take too long.  I understand and
> understood that Daniel is very busy, as are we all.  That hasn't
> changed.
> 
> At this point, lxc is flat out broken on several distributions thanks to
> systemd (pivot_root / MS_SHARE incompatibility).  More recently, Fedora
> 17 updated systemd to a version that breaks lxc where it was working
> previously established Fedora 17 hosts.  What was working is now no
> longer working.  This is very bad.
> 
> Serge and I corresponded over a cleaner answer to the pivot_root problem
> and that may be addressed in a future patch down the road.  The current
> solution may be a bit on the ugly side but it works.  The clean up for
> that and the mount table may wait for another release.
> 
> What is left in our way to cut an alpha3 a beta or an rc?  Seriously?
> This is an alpha.  Can we throw a stake in the ground can cut another
> one?  We just got a report on the -user list where 0.9.0 alpha2 was
> failing on an updated Fedora 17, which is exactly what I would expect
> having experienced it myself.  As soon as we can throw a tag on it,
> someone can throw a bugzilla ticket on it on Fedora (and elsewhere) and
> say THEY are broken and HERE is the fix.
> 
> What's holding us up at this point?
> 
> Regards,
> Mike

Hi,

From one of my previous e-mails to this list:
"""
0.9.0.alpha1 - December
0.9.0.alpha2 - February
0.9.0.rc1    - March
... as many rc as needed to fix any major issue
0.9.0
"""

Now, alpha2 was tagged by mistake, so alpha2 above is really alpha3, but
the rest still makes sense.

The goal for 0.9.0 is to have all the remaining configuration and API
bits in, which we're still pretty far off at the moment as the Bionic
and systemd problems have been taking a significant amount of time from
the people who are working on the remaining bits of the API and on the
configuration rework.

The longer term plan is to have 1.0 release with a stable, versioned
API, so it's crucial that as much of the major work for this is done in 0.9.

As for the systemd situation, the vast majority of our users are
currently using upstart (ubuntu, rhel, oracle), sysvinit (debian, older
rhel, older oracle, ...) or some other init systems (busybox, ...) that
aren't as broken as systemd.
So although I'm clearly happy to review and merge any sensible (and I
stress the sensible part here) fixes for the systemd situation, I have
no plan on altering the release schedule for it or spend any of my own
time on issues created by upstream systemd changes.


I hope this clarifies the situation a bit. Those plans have been
discussed in public on this mailing-list, though on a bunch of separate
e-mail threads (look for the API, 0.9.0alpha1 pull request, ...) so
hopefully the above helps summarize this.

-- 
Stéphane Graber
Ubuntu developer
http://www.ubuntu.comh

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 901 bytes
Desc: OpenPGP digital signature
URL: <http://lists.linuxcontainers.org/pipermail/lxc-devel/attachments/20130119/c93faa91/attachment.pgp>