[Setup] ZODB Replicated Storage experiences?

Patrick Gerken do3ccqrv at googlemail.com
Thu Aug 15 08:29:29 UTC 2013


a customer of ours has been using ZRS before it was open sourced. Two or
three weeks ago we found an issue that forced us to stop using zrs. We are
still waiting for feedback though. Maybe we are just doing something wrong.
Also, we are using an old version, so we do not have blob experience.

We did not notice any performance changes after we switched off
replication. So it seems to be without performance penalties. Both servers
are very close to each other though.

The way WE set up replication, there would be still a long delay between
failovers. I guess with the newer versions and zookeeper support you might
be able to keep you zeo clients running when switching from master to
secondary. I believe you still have to restart your secondary zeo server to
use a different configuration to act as a master. Else there is no writing

I am btw. interested in your Relstorage Setup. Are you using the released
version or the source checkout with the new features. What is the issue
with packing? Does it take long and degrades answer times?

Thanks an best regards,


On Thu, Aug 15, 2013 at 4:48 AM, Héctor Velarde <hector.velarde at gmail.com>wrote:

> as many of you may know by now, ZODB Replicated Storage (ZRS) was released
> as open source software last May:
> https://bitbucket.org/zc/zc.**zrs <https://bitbucket.org/zc/zc.zrs>
> https://pypi.python.org/pypi/**zc.zrs<https://pypi.python.org/pypi/zc.zrs>
> according to its documentation, ZRS "provides database replication for
> ZODB. For each database, a primary storage and one or more secondary
> storages may be defined. The secondary storages will automatically
> replicate data from the primary storage.
> "Replication is superior to back-ups because as long as secondaries are
> running, secondary data is kept updated. In the event of a failure of a
> primary storage, just reconfigure a secondary to be the primary, and it can
> begin handling application requests."
> in the past we have been using RelStorage for high availability and
> replication but we are not extremely happy with it because, at least on our
> experience, on sites with lot of content creation can DB maintenance
> (packing) can be a little bit problematic. besides that, you also have to
> maintain 2 different technologies.
> I just finished reading the documentation and, though fortunately it does
> not have doctest integrated, it also is not very clear, at least not for me.
> I think ZRS could be a killer feature for large, mission critical Plone
> sites.
> does anybody here has any experience with it? any configuration examples?
> how does a secondary storage behaves regarding blobs?
> best regards
> --
> Héctor Velarde
> _______________________________________________
> Setup mailing list
> Setup at lists.plone.org
> https://lists.plone.org/mailman/listinfo/plone-setup
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.plone.org/pipermail/plone-setup/attachments/20130815/f8aab1e1/attachment.html>

More information about the Setup mailing list