Log file Viewer

[13:26:23] <sumanah> I hereby complain about the FASTEN folks not listening to me when I said "you have to post about this publicly in the forum if you want people to attend"

[13:27:13] <sumanah> https://discuss.python.org/t/fasten-workshop-this-week-april-8th/8112/1

[13:42:25] <sumanah> I have now complained. Thank you.

[19:16:46] <fungi> di: in the wake of yesterday's incident, we're still seeing some requests for indices getting served from the bandersnatch fallback (so no data-requires-python metadata), is this something which will eventually clear up on its own or does someone need to poke fastly?

[19:17:31] <fungi> seems like we randomly get copies of the same index with and without metadata for repeated requests

[19:30:48] <PSFSlack> <di> fungi: Shouldn't be happening, but might be possible that the CDN still has some stale responses. If you have specific URLs we can manually purge those

[19:31:39] <fungi> di: thanks! i'll try to collect a list of some, but setuptools was one of them

[19:32:13] <fungi> we had python 2.7 ci jobs upgrading setuptools to 45.0.0 even though that requires python>=3.5

[19:32:27] <fungi> (with a new enough pip to know better)

[19:35:32] <fungi> we already constrain/pin most things explicitly, so the only places we're going to see issues are where we have gaps (particularly around packaging ecosystem stuff like setuptools)

[19:45:50] <fungi> does making purge requests to the suspected stale url still work? if so we can just do that as we run into them

[19:50:51] <PSFSlack> <di> fungi: yep

[19:54:39] <fungi> cool, we'll just do that as we find them

[19:54:44] <fungi> thanks again for all the help!

[20:02:11] <PSFSlack> <di> fungi: we should be able to just purge the entire /simple cache as well