cloudflare / cloudflare/workerd
🐛 Bug Report — SIGABORTs on macOS (JS heap OOM)
- Dominant language
- C++
- Stars
- 8.7k
- Forks
- 739
- Avg merge
- 2d 20h
- Merged PRs (30d)
- 174
Description
About once a day I get a sigabort in one of the workerd instances we use to run our product. It's very frustrating. Happy to help debug.
EDIT: seems like the issue here might be related to [many concurrent requests overwhelming workerd](https://github.com/cloudflare/workerd/issues/3120#issuecomment-2550952443)... that seems like its not good regardles.
```
npx wrangler pages dev --do ROOM_DURABLE_OBJECT=RoomDurableObject@gg-local-gameserver --compatibility-date=2023-09-22 --port=8173 --proxy=6333 --log-level=warn
```
```
✦ ❯ uname -a
Darwin mcbook.local 24.2.0 Darwin Kernel Version 24.2.0: Fri Dec 6 19:01:59 PST 2024; root:xnu-11215.61.5~2/RELEASE_ARM64_T6000 arm64
✦ ❯ npm run wrangler --version
10.2.3
```
```
✘ [ERROR] <--- Last few GCs --->
[99708:0x3a0040000] 11440470 ms: Mark-Compact 1392.6 (1422.9) -> 1390.2
(1423.4) MB, pooled: 6 MB, 163.75 / 0.00 ms (average mu = 0.890, current mu
= 0.724) allocation failure; scavenge might not succeed
[99708:0x3a0040000] 11440671 ms: Mark-Compact 1391.1 (1424.3) -> 1391.1
(1424.3) MB, pooled: 6 MB, 200.42 / 0.00 ms (average mu = 0.779, current mu
= 0.003) allocation failure; scavenge might not succeed
<--- JS stacktrace --->
workerd/jsg/setup.c++:38: fatal: V8 fatal error; location = Reached heap
limit; message = : allocation failed: JavaScript heap out of memory
✘ [ERROR] *** Received signal #6: Abort trap: 6
stack:
```
Contributor guide
Assessment
This issue has not been assessed yet.