DevOps Weekly Log, 2024-02-14

Hello from Friend’s Day (ystävänpäivä) in Finland! Time for another weekly log.

Over the last three days, I worked to fix the search functionality on The root cause was not related to the ongoing migration I’m performing, but fixing the problem took longer than it would have previously, because I still have a lot to learn about how Stackage works. The good news is it’s fixed now. You can read more about it at Hoogle database backing down? · Issue #7314 · commercialhaskell/stackage · GitHub.

Otherwise, I worked on — you guessed it — the Stackage migration itself. Over the weekend we tested pushing the haddocks for new snapshots to the new bucket, and it worked fine. That means it’s time to cut everything over to using the new bucket exclusively, and then do the one-time transfer of all historical haddocks and hoogle databases from the old bucket. That’s all that remains to move Stackage fully into the Haskell Foundation sphere of administration! After that, there will be a number of followup tasks to improve reliability and so on.

It’s been a busy few days, so that’s all the time I have for this log. Next week I will be away, so the next log will be in two weeks.

P.S. I missed the GHC meetings since I was working on Stackage, so there are no GHC issue triage notes this time.


Hello again,

I’ve had limited availability the last two weeks, and furthermore I’ve been blocked on 3rd party issues for most of the work I want to get done. Things are progressing—slowly—but I think I will skip writing a whole separate log this week.

I will also have limited availability for the next two weeks (for different reasons), so there may be one or two more short updates coming.