| daniel added a comment. |
The schema update is not sufficient, we also need code to populate the new column. That code should be triggered by a schema update, but also needs a maintenance script. The script needs to operate in batches of configurable size, and should wait for slaves to catch up between batches.
Besides this, the code that updates the terms table needs to be able to write to the old and the new column at the same time. Similarly, the code that reads from the table needs to fall back to using the old columns in case the new column doesn't exist, or is empty. So we will need a feature switch with three states: numeric ids (legacy), full ids (new), and both (compat).
TASK DETAIL
EMAIL PREFERENCES
To: daniel
Cc: gerritbot, Aklapper, daniel, Lydia_Pintscher, WMDE-leszek, QZanden, D3r1ck01, Izno, Wikidata-bugs, aude, Mbch331
Cc: gerritbot, Aklapper, daniel, Lydia_Pintscher, WMDE-leszek, QZanden, D3r1ck01, Izno, Wikidata-bugs, aude, Mbch331
_______________________________________________ Wikidata-bugs mailing list Wikidata-bugs@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs
