[BUG] Instance/ModuleStoreService: Refreshing store module list failed: fetch failed

[peternewman]

Is this a bug in companion itself or a module?

• I believe this to be a bug in companion

Is there an existing issue for this?

• I have searched the existing issues

Describe the bug

Module list doesn't work through corporate proxy

Steps To Reproduce

1. Configure corporate proxy via http_proxy/https_proxy
2. Load Companion
3. Check log

25.05.21 16:51:15 Instance/ModuleStoreService: Refreshing store module list
25.05.21 16:51:26 Instance/ModuleStoreService: Refreshing store module list failed: fetch failed
25.05.21 16:51:26 Instance/ModuleStoreService: Done refreshing store module list

Expected Behavior

No response

Environment (please complete the following information)

- OS: Debian GNU/Linux 12 (bookworm)
- Browser: Various
- Companion Version: v4.0.0+7969-main-27cc46cfd8

Additional context

No response

[Julusian]

I was so close to having an easy solution, but they havent backported it to node 22 :(
nodejs/node#57165

Hopefully beta 7991 will behave better with NODE_USE_ENV_PROXY=1 also set? This attempts to setup undici (which docs suggest will affect the builtin fetch), to respect the usual HTTP_PROXY and related env vars

[peternewman]

Thanks @Julusian .

However catch-22, I had to monkey-patch /usr/local/src/companionpi/update-prompt/main.js in the same way to allow me to update (have raised bitfocus/companion-pi#40 for this), as I got this originally:

sudo companion-update
Warning: Downgrading to an older version can cause issues with the database not being compatible
You are currently on "4.0.0+7969-main-27cc46cfd8"
✔ What version do you want?  latest beta
node:internal/deps/undici/undici:13484
      Error.captureStackTrace(err);
            ^

TypeError: fetch failed
    at node:internal/deps/undici/undici:13484:13
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async getLatestBuildsForBranch (file:///usr/local/src/companionpi/update-prompt/main.js:20:21)
    at async selectBuildOfType (file:///usr/local/src/companionpi/update-prompt/main.js:47:24) {
  [cause]: ConnectTimeoutError: Connect Timeout Error (attempted address: api.bitfocus.io:443, timeout: 10000ms)
      at onConnectTimeout (node:internal/deps/undici/undici:2602:28)
      at Immediate._onImmediate (node:internal/deps/undici/undici:2583:11)
      at process.processImmediate (node:internal/timers:491:21)
      at process.callbackTrampoline (node:internal/async_hooks:130:17) {
    code: 'UND_ERR_CONNECT_TIMEOUT'
  }
}

Node.js v22.13.0

But I've now hit #3427 so can't test your fix (but the update script will need fixing too...)

[Julusian]

I'm assuming this is working for you now

[peternewman]

Yes thanks @Julusian . Sorry I ran out of time to test beforehand.

I had to add the following to my /etc/systemd/system/companion.service file:

Environment=NODE_USE_ENV_PROXY=1
Environment=ftp_proxy="http://example.com:80"
Environment=http_proxy="http://example.com:80"
Environment=https_proxy="http://example.com:80"

I guess as a proxy user/sufferer, I'd wonder if we should default to setting NODE_USE_ENV_PROXY in the shipped SystemD file, as I assume it doesn't cause any harm to non-proxy users?

[Julusian]

I'm hesitant to do so for now, purely because I don't know the implications of it and nodejs themselves are unsure if it will cause issues for some users.