Manuel Pöter

Wraparound NextCompactionIndex in LevelCompactionBuilder

4

There can be situations where the picker is asked for a compaction twice for the same version. In such a scenario the first call might not be able to find...

[3.9] Fix potential deadlock in rocksdb compaction.

1

### Scope & Purpose Backport for #16597 There can be situations where the picker is asked for a compaction twice for the same version. In such a scenario the first...

[Fast concurrent lock-free binary search trees](https://dl.acm.org/doi/abs/10.1145/2555243.2555256)

[Hyaline: Fast and Transparent Lock-Free Memory Reclamation](https://arxiv.org/abs/1905.07903)

- [x] add Clang builds - [x] add Windows builds - [ ] add Mac builds - [ ] add ARM builds - [ ] add Power builds

For various reasons we run all our queries (and mutations) via WebSocket and found that all requests are handled sequentially in `Absinthe.Phoenix.Channel`. It would be great if we could execute...

Feature proposal: define and respect limited number of available resources

5

##### Issue Type - Feature proposal ##### Summary Allow to define a fixed number of available resources and respect this limit when scheduling jobs that require this resource. E.g. lets...

Feature: avoid write-write conflict errors for single document updates

38

When using RocksDB, two concurrent update operations on the same document can result in a write-write conflict (error-code 1200). This type of error often causes confusion among users (this is...

### Scope & Purpose These tests rely on specific behavior of the old replication protocol, so it makes so sense to run them with replication 2. - [x] :hankey: Bugfix