Kirill Bulatov
Kirill Bulatov
Also see https://github.com/neondatabase/neon/pull/2560#issuecomment-1269512883 1. Regression tests fail: ``` FAILED test_runner/regress/test_remote_storage.py::test_remote_storage_backup_and_restore[RemoteStorageKind.REAL_S3] [18087](https://github.com/neondatabase/neon/actions/runs/3191604676/jobs/5215850177#step:4:18106) FAILED test_runner/regress/test_tenants.py::test_pageserver_with_empty_tenants[RemoteStorageKind.REAL_S3] [18088](https://github.com/neondatabase/neon/actions/runs/3191604676/jobs/5215850177#step:4:18107) FAILED test_runner/regress/test_tenants_with_remote_storage.py::test_tenants_many[RemoteStorageKind.REAL_S3] [18089](https://github.com/neondatabase/neon/actions/runs/3191604676/jobs/5215850177#step:4:18108) FAILED test_runner/regress/test_wal_acceptor.py::test_wal_backup[RemoteStorageKind.REAL_S3] [18090](https://github.com/neondatabase/neon/actions/runs/3191604676/jobs/5215850177#step:4:18109) FAILED test_runner/regress/test_wal_acceptor.py::test_s3_wal_replay[RemoteStorageKind.REAL_S3] [18091](https://github.com/neondatabase/neon/actions/runs/3191604676/jobs/5215850177#step:4:18110) ``` 2. e2e tests trigger fails: ```...
Yes, that's the reason for the most of it. If we won't be implementing a solution to the creds soon, we can meanwhile detect if that's an external PR and...
> 1. At pageserver startup, if some files are missing locally, but are present in S3, were they downloaded? With this PR, they are not. They probably should be. Yes,...
Not quite sure what does the `...` in the end mean, but pretty sure we're not merging this PR, especially now, when big parts of it are merged and https://github.com/neondatabase/neon/pull/2433...
https://github.com/neondatabase/neon/pull/2426 and https://github.com/neondatabase/neon/pull/2489 implemented atomic creation of timeline and tenant files (including their directories). Tenants are created through the tmp directory as proposed here, timelines are currently created using a...
Most probably, those > Processing 1 events are the other ones sent by safekeepers, related to its WAL offload. One could check the events with `etcdctl --endpoints=etcd-release.local:2379 watch --prefix neon/9a545ef7f96d4e3ba3e60823c5e36e0f/33760f91928b64a7c62dd431507b92ab`...
That's a great finding, thank you for looking into that. TBH, this solution feels like a stub to me, fixing the issue with a non-obvious way, but have no better...
Great ideas, thank you for adding those. 1. I lack the good knowledge on gc, compaction and page materialization, so not sure can suggest the right order myself. Yet I...
Agreed multiple times, that we should merge it deal something with this, presumably merge it. Feel free to revert the commit, if this was rushed.
Closing as no one had reclaimed it and it is pretty vague at the current point. There are nice comments here, would be good to reuse on the next iteration.