Release 0.7.4
Arjan Scherpenisse
Yesterday zotonic 0.7.4 was released. This release is a minor bugfix
release which, amongst others, fixes the issue that a clean install of
the "blog" skeleton site shows a 404 page instead of the home page.
As always, it can be downloaded from Google code:
http://code.google.com/p/zotonic/downloads/list
Note that most current development of Zotonic is done on the "Zynamo"
branch: this is the branch for the "Elastic Zotonic" project you heard
about: an effort to make Zotonic work in a distributed fashion over any
number of nodes. In the coming weeks/month you'll hear more about this
effort.
kind regards,
the Zotonic team
Konstantin Nikiforov
before. I want to share some own expirience about zotonic clustering.
I have a project.
It is unreleased right now, but it is big-data oriented web-service.
I've written several data models.
These models utilizing Riak[1] as storage backend.
So, zotonic became to an easy framework and frontend for this project.
Postgresql, CMS and some other parts/modules are unused.
I've found some main problems of clustering of zotonic. In short:
1. Sessions and page_sessions has minimum clustering abilities.
2. PostgreSQL... It is good RDBMS, but it is totally unscalable.
So i try to avoid to use it in big-data-related projects.
Some minimal and/or compile-time db-backend abstraction will be
welcome).
Some current crunches for using riak:
1. To use full text search in riak_search i have wrote libstemmer
bindings to erlang and use it as text analyze factory.
libstemmer[2] is a word stemmer used in postgresql.
2. Complex sql requests are rewritten using map-reduce javascript.
[1] http://wiki.basho.com/Riak.html
[2] http://snowball.tartarus.org/download.php
ps: This message is not a call for using or not using riak as
storage for zotonic :) Only some my expirience.
В Wed, 11 Jan 2012 10:57:25 +0100
Arjan Scherpenisse <ar...@scherpenisse.net> пишет:
Dmitrii Dimandt
Arjan Scherpenisse
On 01/11/2012 12:26 PM, Dmitrii Dimandt wrote:
> Zotonic so far still has quite a few hardcoded sql queries around it's
> database. If these can be removed in favor of using calls to z_db, then
> it's not impossible to change z_db to use some other database.
That's exactly what we are doing. Abstracting away all SQL calls so the
backend can be swapped.
>
> IMO, I haven't looked at Zotonic internals for quite some time now
>
> 2012/1/11 Konstantin Nikiforov <hell...@gmail.com
> <mailto:hell...@gmail.com>>
>
> It is good to hear about "zynamo" branch. I've not heard about it
> before. I want to share some own expirience about zotonic clustering.
>
> I have a project.
> It is unreleased right now, but it is big-data oriented web-service.
> I've written several data models.
> These models utilizing Riak[1] as storage backend.
> So, zotonic became to an easy framework and frontend for this project.
> Postgresql, CMS and some other parts/modules are unused.
>
> I've found some main problems of clustering of zotonic. In short:
>
> 1. Sessions and page_sessions has minimum clustering abilities.
> 2. PostgreSQL... It is good RDBMS, but it is totally unscalable.
> So i try to avoid to use it in big-data-related projects.
> Some minimal and/or compile-time db-backend abstraction will be
> welcome).
>
> Some current crunches for using riak:
>
> 1. To use full text search in riak_search i have wrote libstemmer
> bindings to erlang and use it as text analyze factory.
> libstemmer[2] is a word stemmer used in postgresql.
>
> 2. Complex sql requests are rewritten using map-reduce javascript.
>
> [1] http://wiki.basho.com/Riak.html
> [2] http://snowball.tartarus.org/download.php
>
> ps: This message is not a call for using or not using riak as
> storage for zotonic :) Only some my expirience.
>
>
> О©╫ Wed, 11 Jan 2012 10:57:25 +0100
> Arjan Scherpenisse <ar...@scherpenisse.net
> <mailto:ar...@scherpenisse.net>> О©╫О©╫О©╫О©╫О©╫:
Bo
can you speculate on the eventual release date of the "Elastic
Zotonic".
On Jan 11, 1:34 pm, Arjan Scherpenisse <ar...@scherpenisse.net> wrote:
> On 01/11/2012 12:26 PM, Dmitrii Dimandt wrote:
>
> > Zotonic so far still has quite a few hardcoded sql queries around it's
> > database. If these can be removed in favor of using calls to z_db, then
> > it's not impossible to change z_db to use some other database.
>
> That's exactly what we are doing. Abstracting away all SQL calls so the
> backend can be swapped.
>
>
>
>
>
>
>
>
>
> > IMO, I haven't looked at Zotonic internals for quite some time now
>
> > <mailto:hellla...@gmail.com>>