Heikki Linnakangas e4d90fe0bb Add timeouts to WAL redo, to prevent it from getting stuck.
It is getting stuck at least in spgist index currently. Not sure why,
that needs to be investigated, but having a timeout is a good idea anyway.
2021-03-29 15:59:28 +03:00
2021-03-26 10:54:25 -07:00
2021-03-26 10:54:25 -07:00
2021-03-26 10:54:25 -07:00
2021-03-26 10:54:25 -07:00
2021-03-26 10:54:25 -07:00
2021-03-26 10:54:25 -07:00
2021-03-29 15:59:28 +03:00
2021-03-29 15:59:28 +03:00
2021-03-29 15:59:28 +03:00
2021-03-29 15:59:28 +03:00
2021-03-26 10:54:25 -07:00

zenith

This is the top level repo containing the following submodules

  1. vendor/postgres - upstream postgres code
  • zenith branch contains our changes
  • These changes are relatively small and will be either upstreamed or moved to an extension
  1. pageserver - the code that deals with buffer pages disaggregated from compute nodes.

  2. walkeeper - the code that takes the WAL from postgres and signals to postgres when it's safe to consider a transaction committed.

  3. consensus - Implement distributed consensus between compute and storage nodes.

  4. cli - Tooling used to have a friendly workflow that allows migration to/from zenith storage system and cloud provider's object storage services.

Recommended workflow:

$ git clone https://github.com/libzenith/zenith
$ git submodule update --init --recursive
$ cargo build
$ cargo test
Description
Neon: Serverless Postgres. We separated storage and compute to offer autoscaling, code-like database branching, and scale to zero.
Readme Apache-2.0 133 MiB
Languages
Rust 73.5%
Python 19.4%
C 5.2%
Dockerfile 0.8%
Shell 0.3%
Other 0.8%