On Sep 25, 2008, at 11:57 AM, Raoul Bhatia [IPAX] wrote:
hi,
i am now given the choice of using pacemaker 0.7/1.0 or using
pacemaker 0.6.x in a production environment. it's a two-node
cluster used for webhosting services (apache, mysql, ...)
using primitives (e.g. mysql-server), clones (e.g. apache) and
multistate resources (e.g. drbd8).
since the 0.7 series now seems to get more attention, more and more
bugs and/or regressions are coming to the surface.
so i was wondering what your thoughts - and especially your's, andrew,
lars, dejan - are. is 0.7/1.0 ready for production use in the very
near
future?
cluster deployment will be in roughly two to four weeks.
I would suggest 0.7/1.0 (but then of course I would, wouldn't I ;-)
The feature-set is frozen and I began release testing yesterday.
So far my _8_ node cluster (40+ resources) has gotten through almost
1000 iterations of CTS without a major incident.
Which makes me quite confident in the core functionality carried over
from 0.6.
The most likely places for bugs are the CLI tools (which are easily
fixed) or the newer features which haven't been bashed on as much.
And certainly if you start testing now, you can be sure 1.0.0 will
work great in your environment :-)
_______________________________________________
Pacemaker mailing list
[email protected]
http://list.clusterlabs.org/mailman/listinfo/pacemaker
