Alex Chi Z.

> and rely on the single check in Timeline::put, which all writes will hit eventually it seems that Timeline::put has `#[cfg(test)]`, it's only used in test cases? 🤣

One interesting observation is that solely doing create tablespace does not reproduce the invalid key error. Next step is to understand what will happen if an invalid key is written...

https://github.com/neondatabase/neon/pull/8648 now rejects keys on the write path while https://github.com/neondatabase/neon/pull/8657 disables the statement. Closing the issue for now. Follow-up is how does safekeeper handle pageserver key rejections, currently this is...

the Postgres submodule seems to have 200+ files updated...? probably you're not on the right branch :(

both of us are on vacation last week unfortunately :( cc @danieltprice @clarkbw what's your take here? does it look like a good direction to smplify the readme file and...

Sorry I was busy in the previous month -- I'll keep this open until I have time to investigate. Thanks for the report :)

I'll revisit this issue before closing it, thanks for filing anyways :)

I pushed the changes to a wrong branch 🤣 will piggyback my commits to other patches.

related fix: https://github.com/neondatabase/neon/commit/db24ba95d1991a90c50d25f322eb07435cf31936

for the project in staging, decide to extend start timeout so that at least it could start