[Cubicweb] Making changes to cubes

Nicolas Chauvat nicolas.chauvat at logilab.fr
Sun Dec 12 22:50:34 CET 2010


On Sun, Dec 12, 2010 at 01:16:29AM +0000, Fletcher, John P wrote:
> Thank you for your help in this and the other messages.
> I had not appreciated that modification of the schema meant that there had also to be a migration.
> I have been able to use the information you gave to get things working and also see the potential for the systems.
> I am particularly liking the way in which "see also" works in a reciprocal way to enrich the data.
> I like also that when I change the structure I can see that reflected in the schema diagram.

Keep in mind that we use CubicWeb for large systems in
production. Which means that we want everything, including schema
changes and migration, to be reproducible, hence written down. We
try things out while developing, then get the result, copy the
production database (db-dump, db-restore or db-copy), then run the
migration script, then fix the mistakes and re-run the migration from
a fresh restore, etc.

In our process, we end up publishing a Debian package and installing
it on the production system, then typing:

   cubicweb-ctl upgrade myprodinstance

and everything must migrate and restart without a glitch.

It does work :)

-- 
Nicolas Chauvat

logilab.fr - services en informatique scientifique et gestion de connaissances  


More information about the Cubicweb mailing list