[imp] IMP/cvs stability

[Eric Rostetter]

Quoting "\"Gonzalo G. Agulló\"" <agullo@comtenidos.com>:

> Hi,
> 
> Currently running IMP 3.1 successfully, but I'm missing improvements and
> fixes only available from cvs. I wonder how stable cvs branches are.

It varies.  The Horde/IMP CVS branch had been very stable for some time now
(last several months) but lately (in the last week or so) Chuck kind of
stirred things up.  So right now, if you use kronolith/nag/mnemo I would say
it is rather unstable, otherwise (without those) it is rather stable.

I'd been using it and upgrading regularly without issue for months now, but
had to stop the upgrades last week after the latest "stirring up" of things ;)

> Are they reasonably ready for production systems if they're not critical
> or not with high load?

I think Horde/IMP are, but other modules may not be.

> I have a few IMP balanced frontends, sharing mysql at a backend server
> for storage. If I upgrade some individual server for evaluation before
> setting the whole frontline servers with new version, might I break imp
> tables consistency? Will different branches coexist fine while sharing
> the db backend?

Not neccessarily.  The nag/mnemo/kronolith databases have changed such as
to break backwards compatibility.  But a lot of this is actually backwards
compatibility with CVS HEAD itself, and may or may not affect your old 
release database data.  

I think the IMP filters storage also will break backwards compatibility
(when you run the new version, it will autoconvert to the new format, but
then I think the old version would no longer work with the filters).

So I wouldn't recommend running old and new side by side, if you use any
of the above features.  And there could be more I'm missing, but I think
those are the only ones.  If you don't use the above features, then it
may work well running them side-by-side off one database.

> Thnx

-- 
Eric Rostetter
The Department of Physics
The University of Texas at Austin

Why get even? Get odd!