Closed rebeccacremona closed 1 year ago
I'm not sure where the
I think if they are sent to stdout when we run manage.py migrate
, we'll be able to see them by looking at the output of the perma-migrate state run -- and we can confirm this when we deploy to stage.
Merging #3302 (5077afd) into develop (bbbec52) will decrease coverage by
0.11%
. The diff coverage isn/a
.:exclamation: Current head 5077afd differs from pull request most recent head 2848433. Consider uploading reports for the commit 2848433 to get more accurate results
@@ Coverage Diff @@
## develop #3302 +/- ##
===========================================
- Coverage 73.64% 73.54% -0.11%
===========================================
Files 55 55
Lines 6641 6641
===========================================
- Hits 4891 4884 -7
- Misses 1750 1757 +7
see 1 file with indirect coverage changes
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.
Now that https://github.com/harvard-lil/perma/issues/3296 is fixed going forward, this PR fixes our data, to the extent that we can: it looks in the axes login records and finds the most recent one for each user, and puts that date into
last_login
. There's a gap of a few months where we just don't have data... but c'est la vie.Run locally against ~80K records in a prod-like database, this took 1.5s... so I decided to do it cleanly as a data migration. So, if it takes a long time, that means we'll have a slower deployment than usual... but I think even if it ends up being much slower, adding a minute or two while we are in maintenance mode is perfectly fine.
I'm not sure where the
print
s will be visible, because this will happen as part of a Salt orchestration. @bensteinberg do you know? if you would recommend using thelogger
instead of printing for any reason, happy to switch.