In T380846#10497204, @jsn.sherman wrote:

In T380846#10485408, @jsn.sherman wrote:

Leaving in QA for now. I was still able to create a DB timeout error when doing a title query with the 90 day limit; the initial query threw the error, but subsequent queries worked, presumably due to caching; We should retest again with a cold cache at 60 and 30 days to verify if the error is due to the longer timespan. We realized that we didn't actually have a known error free result before expanding the time range.

It is possible to perform title queries that result in a database timeout, however, the behavior is the same at 60 and 30 days. This means the issue was already there and is not a regression from this task. Will file a bug report and mention this task before closing.

Steps to replicate the issue (include links if applicable):

• go to https://en.wikipedia.org/wiki/Special:Nuke
• run a query with a % on the left-hand side

What happens?:
Database error

Error

• mwversion: 1.44.0-wmf.13
• reqId: 68cd3809-94bb-4301-8d8e-44bd9ab02b89
• Find reqId in Logstash

Error 1969 from MediaWiki\Extension\Nuke\SpecialNuke::getNewPages, {error} {sql1line} {db_server}

Impact

certain enwiki Special:Nuke page title queries can't be run because they regularly exceed the query execution limit

Notes

You must have the massdelete privilege to view the page, but the query can be examined in logstash and run from a stathost for testing