Zhanxiang (Patrick) Huang

compaction: limit ongoing compaction task per compactor

1

**Is your feature request related to a problem? Please describe.** When running the benchmark described in #4365 with 3-cn 1-compactor setting, I noticed that there can be more than 100...

bug: MV gets stuck on node failure

2

**Describe the bug** I did a simple benchmark with materialized source + two-way join in a 3-node setting. After running the benchmark for ~2hrs, one compute node's heartbeat times out...

storage: compactor cannot recover from ObjectIoError

7

**Describe the bug** After running a simple materialized source workload for a while, compaction fails due to ObjectStoreIOError in compactor and compactor fails to re-register itself in meta causing compaction...

storage: No compaction triggered on an imbalanced LSM tree when the cluster is idle

4

**Describe the bug** When the cluster is idle (i.e. source is stopped), we will not trigger compaction on an imbalanced LSM tree after compactor recovers. This happens when I stopped...

fix(frontend): fix unpin_snapshot_before in HummockSnapshotManager

5

I hereby agree to the terms of the [Singularity Data, Inc. Contributor License Agreement](https://gist.github.com/skyzh/0663682a70b0edde7ae991492f2314cb#file-s9y_cla). ## What's changed and what's your intention? Prior to this PR, frontend will never unpin snapshot...

### Is your feature request related to a problem? Please describe. See https://github.com/risingwavelabs/risingwave/pull/5515#issuecomment-1255779170 ### Describe the solution you'd like _No response_ ### Describe alternatives you've considered _No response_ ### Additional...

metrics(compaction): Huge compaction cancelled count showed in grafana

9

### Describe the bug I saw a huge number of cancelled compaction task when running nexmark_q7, which looks abonormal. However, source throughput is not affected. If this is expected, I...

compaction: improve compaction trigger

19

### Is your feature request related to a problem? Please describe. Currently we have the following compaction triggers: - On commit_epoch. It happens when the checkpoint process completes. - On...

storage: maintain L0 SST index in compute nodes

2

## Background In our system, the reader of the internal states is always the writer of the state. Currently, the storage LSM read path of internal states is unaware of...

I hereby agree to the terms of the [RisingWave Labs, Inc. Contributor License Agreement](https://gist.github.com/TennyZhuang/f00be7f16996ea48effb049aa7be4d66#file-rw_cla). ## What's changed and what's your intention? ## Checklist - [ ] I have written necessary...