Page MenuHomePhabricator

Import some content on beta.wmflabs.org wikis
Closed, DeclinedPublic

Description

When trying to use beta.wmflabs.org to test (https://www.mediawiki.org/wiki/Thread:Talk:PDF_rendering/Test_instance_and_bug_triage ), I noticed that only simple.wiki got content imported and he.wiki has a partial import (3k pages).

All articles and templates (if not all pages), current version, need to be imported on the following existing wikis: ar de en eo fa hi ja ko ru sq uk zh. Or you could leave en aside as it's so big and we already have simple in English anyway.

https://meta.wikimedia.org/wiki/Data_dumps/Tools_for_importing can be of use. Needs to be done server-side because Special:Import timeouts after few dozens MB.


Version: unspecified
Severity: enhancement
URL: http://dumps.wikimedia.org/

Details

Reference
bz66402

Event Timeline

bzimport raised the priority of this task from to Low.Nov 22 2014, 3:13 AM
bzimport set Reference to bz66402.
bzimport added a subscriber: Unknown Object (MLST).

The beta cluster is not powerful enough to be a close copy of the production wikis. En and de would definitely not fit on the small database server.

Moreover, I am not sure there is a point in having the millions of pages imported if they are never going to be accessed / used.

If there is a specific need, one can export pages they are interested in and manually import them on the wikis.

So unless there is a specific testing needs to import all the pages, I am willing to discard this request :D

(In reply to Antoine "hashar" Musso from comment #1)

If there is a specific need, one can export pages they are interested in and
manually import them on the wikis.

This is disproved by reality. The one time I had to import something on a wiki, timeouts disallowed me to. Do you have proposals on how to integrate page + templates import of test cases *during* a bug triage.

If you want to go through https://etherpad.wikimedia.org/p/BugTriage-mwlib and determine all the test cases that need to be produced beforehand, feel free to.

"Import some content" sounds valid but blocked by beta cluster not powerful enough (comment 1).

Nemo: What is the criteria for "ar de en eo fa hi ja ko ru sq uk zh"? Could that list be shortened && a partial import of some of their pages?

(In reply to Andre Klapper from comment #3)

Nemo: What is the criteria for "ar de en eo fa hi ja ko ru sq uk zh"?

Just the list of existing and empty wikis.

Could
that list be shortened

I have no idea how the wikis to create were chosen, i.e. what are the wikis "we care about" in beta.

&& a partial import of some of their pages?

There is no secure recursive import feature, so at least all templates and modules would need to be imported anyway. Maybe ns0 pages could be randomly selected, but I'm not going to engage in XML surgery.

There is no need to import the whole prod wiki on beta. If there is a need for a bulk import of some contents, we can use the command line utilities to export from prod and import on beta.

Hence I am closing this bug.