John Spray
John Spray
Just noticed this is affecting test_forward_compatibility too, so would be good to fix it soon.
Can we reconcile this with https://github.com/neondatabase/neon/issues/6345? (I don't mind replacing that issue, just want to avoid duplication)
@VladLazar is this issue still relevant now that we're using vectored get by default everywhere, or is the LSN->timestamp lookup still doing its own thing?
Thinking of extra safety measures: we might in future like to carry an HTTP header on controller requests to pageservers, which would change for new leaders, so that pageservers could...
Does this become simpler if we implement it by finding the index from a point in time, and then just undelete all the layers that are referenced by the index?...
This might still be failing
> Storage team: any concerns from generating more churn on the single "checkpoint" key-value pair? Can we get a test on this PR that reproduces the many-transactions case that the...
It's interesting that these occur on release mode tests. I wouldn't be surprised on debug, but for release mode tests to take 30s+ to create a timeline indicates maybe that...
We should at least figure out which part was slow (controller, pageserver) and check the metrics for the runners.
> Internally retry the attachment in the Pageserver. +1 for this - it was always wrong to give up after some retries to S3 -- we don't stop wanting the...