clarify write batching is crash-safe (client blocks until durable)

Chris Lu
2026-03-09 19:55:27 -07:00
parent 300d360f10
commit 52226341de
+1 -1
@@ -175,7 +175,7 @@ The `redb` backend is Rust-only. It provides crash-safe disk-backed needle maps
| `/metrics` endpoint | On admin/metrics port | On admin (debug) port | Both serve on admin port, not the main volume port |
| Disk monitoring | `sysinfo` package | `sysinfo` crate + `libc` | |
| Volume preallocation | `fallocate` on Linux | `fallocate` on Linux via `libc` | Identical behavior |
| Write batching | Synchronous per-request | Async batched write queue (128 batch, mpsc) | Rust batches concurrent writes for higher throughput |
| Write batching | Synchronous per-request | Batched write queue (up to 128, mpsc) | Rust groups concurrent writes to reduce lock acquisitions. Each client still blocks until its write is durable — no crash-safety difference |
| Streaming reads | Files > 1MB | Files > 1MB via `spawn_blocking` | Same threshold, async I/O |
### Extra CLI Flags (Rust only)