I don't know much about how the dump is generated - is it just excluding anything with a custom policy? That shouldn't break - policy is still stored the same way it's just user interface stuff that's changing for the most part.

@JAufrecht: Upgrade happened. When you have time, please report if the dump broke. Or not. :) Thanks.

It seems to be fine this morning, but James points out that the migration of the points field might have happened after the dump? Let's give it one more day to be sure.

@JAufrecht: T127524#2050513 implies this task is invalid (nothing to fix)?

@Aklapper: @mmodell had to fix it with https://gerrit.wikimedia.org/r/#/c/272320/

Uh! Thanks, wasn't aware!

When I first checked it, I misread the dates by a few hours and I was still working pre-dump. It was broken from Thu to Sun, ran successfully Sunday. Let me spend another day or two checking before declaring this okay.

Some data is missing from the new Phlogiston reports; will need to investigate further.

Could you share which data (if you already have an overview)?

can this be closed now?

It's not active. I find it helpful to coordinate all of the related issues when a new one comes up - should I re-open this with a new sub-task if that happens?

It's ok, I'll just unsubscribe. I've been monitoring this to be sure that the issue was resolved.

@JAufrecht: Is this still an issue / should this task remain open? Or do things work more reliably in these days?

It's been stable since August, after the db upgrade. However, it is useful to have all of the outage tasks linked together, e.g. via this parent task. Some options:

1. leave as is. I assume having this task open is annoying, though.
2. close this tracking task, and then I'll re-open it if there are new children
3. Instead of using a parent task, use some other method such as a tag or column in a tag to relate these tasks to each other.

Removing Phabricator as there's nothing to do there.

Should we do #2 instead, leave this as Phabricator but close it?