Sementation fault since node-sass 4.7.1

[Vinnl]

• npm version (npm -v): 3.10.10
• Yarn version: 1.3.2 (I'm using Yarn)
• Node version (node -v): 6.11.2
• Node Process (node -p process.versions):

{ http_parser: '2.7.0',
  node: '6.11.2',
  v8: '5.1.281.103',
  uv: '1.11.0',
  zlib: '1.2.11',
  ares: '1.10.1-DEV',
  icu: '58.2',
  modules: '48',
  openssl: '1.0.2l' }

• Node Platform (node -p process.platform): linux
• Node architecture (node -p process.arch): x64
• node-sass version (node -p "require('node-sass').info"): 4.7.1 (manually inserted - I didn't get this command to run in Docker)
• npm node-sass versions (npm ls node-sass): I couldn't get this to run in Docker either...

I'm running my project in a Docker image. However, this recently started running into segmentation faults, and some git bisecting taught me that this started occurring after an upgrade of node-sass. I then further narrowed it down to the segfault not occurring with node-sass 4.6.1, but occurring from 4.7.1 onwards. This, of course, makes PR #2128 suspect. The issue sounds similar to #2031, but since that predates 4.7.1, I figured it was probably a different one. However, most of that issue went over my head.

Unfortunately, I don't know much more about segfaults than that they are related to memory management. Thus, if there's any additional things I can try or information I can gather, let me know.

I'm running inside the node:6-alpine Docker image (using this Dockerfile.

This is, I think, the relevant but not very informative output:

web_1       | [2] Project is running at http://0.0.0.0:8000/
web_1       | [2] webpack output is served from /
web_1       | [2] 404s will fallback to /index.html
web_1       | [2] Segmentation fault (core dumped)
web_1       | [2] error Command failed with exit code 139.

At this point my Webpack server wasn't actually running, and I Ctrl+C'd out.

[Vinnl]

Update: I also tried it with the node:8-alpine Docker image, and that works fine. (Unfortunately, I have to use 6 for now...)

[tom-pryor]

Yeah, we ran into the same problem. Fortunately we were able to use node 8 which doesn't have this problem.

[saper]

#2031 is not a problem with node-sass, it's the problem with anything musl, including alpine.

Can you publish something to reproduce this?

Can you reproduce the problem and get a backtrace like this #2031 (comment) ?

@Vinnl did you try top20/node:8-alpine as patched by @jubel-han in #2031 (comment) - or something similar for node 6 (stackfix should work for node 6 as well) ?

[nschonni]

Between 6 and 8, the Node docker images updated the version of Apline used

[Vinnl]

@saper I don't think I'll be able to make a minimally reproducible version (unless there already is a basic node-sass project somewhere that I could set up in a Docker image?), but it should be reproducible by cloning this repo and then checking out commit 03d9aebe.

Unfortunately, the backtrace/stackfix goes a bit over my head, so that doesn't look like I would be able to reproduce it, and since it works in the regular Node 8 image, the one already made probably works as well.

@nschonni makes an interesting point: the Node 6 image uses Alpine 3.4, and the Node 8 image uses Alpine 3.6.

[tom-pryor]

I don't think the version of Alpine matters. Our base image that we add node to is still using Alpine 3.4 - we installed Node 8 by specifying the edge repo in the apk add command.

For reference:

RUN apk add \
    --repository http://dl-cdn.alpinelinux.org/alpine/edge/main \
    --no-cache \
    nodejs

[cdcme]

I experienced this issue, as well. My fix was to also to use node 8. I also force an npm rebuild of node-sass in that image. This resolves the issue for me.

[saper]

This is the same as #2031. Do not use Alpine for node < 8.