[ckan-discuss] Architecture Best Practice

David Read david.read at hackneyworkshop.com
Tue Apr 30 22:45:08 BST 2013


Marina,

Great to hear about your project - sounds excellent.

Looking after one CKAN server is a bit of work, so your consideration
of running 5 to 10 of these things concerns me. Unless they really are
almost identical configurations. You know the sort of thing -
monitoring your backups are running nightly and they aren't using up
all your disk space, then you find a CKAN bug and want to install the
patch, and then you need to change the caching settings or migrate the
database and install a new plugin. To do a good job with these sorts
of issues coming up, you'll really need something like Puppet to keep
track of the configurations. So if you can do it all with one CKAN
instance then that might save on maintenance work.

300 to 1200 extra fields would worry me somewhat from the performance
point of view, although my experience doesn't include the SOLR caching
tricks that CKAN 2.0+ has. All the same I think it is well worth doing
some tests with a schema this big and a suitable number of datasets
(stuffed with test data, even if all the same) as soon as you can, to
check that CKAN runs acceptably.

David


On 30 April 2013 13:03, Sean Hammond <sean.hammond at okfn.org> wrote:
>> Speaking of validation, is there any documentation I'm missing on
>> validators.py? I couldn't quite tell how those validations are currently
>> being called. (For example I tried to work backwards in the code how
>> default CKAN stops you if you try to enter a name under 2 characters for a
>> dataset, but I couldn't clearly find where it was being called.)
>
> This wiki page might be relevant:
>
> https://github.com/okfn/ckan/wiki/Converters-and-Validators-in-navl-Schemas
>
> The validators are called by the validate() function (in
> ckan/lib/navl/dictization_functions.py). Each action function in
> ckan/logic/action/*.py should call validate() at some point, passing in
> the appropriate schema.
>
> _______________________________________________
> ckan-discuss mailing list
> ckan-discuss at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/ckan-discuss
> Unsubscribe: http://lists.okfn.org/mailman/options/ckan-discuss



More information about the ckan-discuss mailing list