build errors

[ippocratis]

trying to build on a raspbbery 400 running manjaro arm connected to eth0

running sudo docker-compose -f ./server/docker-compose.yml up -d
gives
server error
and

[2/4] Fetching packages...

info There appears to be trouble with your network connection. Retrying...
error An unexpected error occurred: "https://registry.yarnpkg.com/class-validator/-/class-validator-0.13.2.tgz: ESOCKETTIMEDOUT".
info If you think this is a bug, please open a bug report with the information provided in "/usr/src/app/yarn-error.log".
info Visit https://yarnpkg.com/en/docs/cli/install for documentation about this command.
ls 1 error occurred:
	* Status: The command '/bin/sh -c yarn install' returned a non-zero code: 1, Code: 1

[alextran1502]

I saw this on my build for ARM as well, I will investigate this.

[alextran1502]

I updated the Dockerfile to use npm instead of yarn, can you try again?

[ippocratis]

recloned the repo
and re-run docker-compose
also failed to build
had a server error again
and a debconf: delaying package configuration, since apt-utils is not installed error on Step 6/11
on Step 7/11 : RUN npm install

 ---> Running in c2ed2180bd63
npm WARN old lockfile 
npm WARN old lockfile The package-lock.json file was created with an old version of npm,
npm WARN old lockfile so supplemental metadata must be fetched from the registry.
npm WARN old lockfile 
npm WARN old lockfile This is a one-time fix-up, please be patient...
npm WARN old lockfile 
npm notice 
npm notice New minor version of npm available! 8.1.2 -> 8.4.1
npm notice Changelog: <https://github.com/npm/cli/releases/tag/v8.4.1>
npm notice Run `npm install -g npm@8.4.1` to update!
npm notice 
npm ERR! code 1
npm ERR! path /usr/src/app/node_modules/@tensorflow/tfjs-node
npm ERR! command failed
npm ERR! command sh -c node scripts/install.js
npm ERR! CPU-linux-3.13.0.tar.gz
npm ERR! * Downloading libtensorflow
npm ERR! /usr/src/app/node_modules/@tensorflow/tfjs-node/scripts/install.js:106
npm ERR!     throw new Error(`Unsupported system: ${libType}-${platform}-${os.arch()}`);
npm ERR!           ^
npm ERR! 
npm ERR! Error: Unsupported system: cpu-linux-arm64
npm ERR!     at getPlatformLibtensorflowUri (/usr/src/app/node_modules/@tensorflow/tfjs-node/scripts/install.js:106:11)
npm ERR!     at downloadLibtensorflow (/usr/src/app/node_modules/@tensorflow/tfjs-node/scripts/install.js:139:15)
npm ERR!     at async run (/usr/src/app/node_modules/@tensorflow/tfjs-node/scripts/install.js:208:5)

npm ERR! A complete log of this run can be found in:
npm ERR!     /root/.npm/_logs/2022-02-07T22_50_27_449Z-debug.log
1 error occurred:
	* Status: The command '/bin/sh -c npm install' returned a non-zero code: 1, Code: 1

i think apt-utils and old npm versions are not critical?
but unsuported arch for tensorflow install script is?

[alextran1502]

Yes, Tensor Flow 2 right now supports 32bit system on Raspberry Pi but they don't have plan for 64bit yet. I will keep you update on this issue

[ippocratis]

i could try something like

https://github.com/photoprism/photoprism/blob/adeb18dc443f83d0d576e6f236587e6b277465d4/docker/scripts/install-tensorflow.sh

or

https://github.com/bitsy-ai/tensorflow-arm-bin

with prebuild tensorflow binaries for arm64?

[alextran1502]

You can try but I am afraid it won't work since the binary is built when installing the npm packages for the server. I hope they will be adding support for arm64 soon.

[ippocratis]

You can try but I am afraid it won't work since the binary is built when installing the npm packages for the server. I hope they will be adding support for arm64 soon.

I see
So
Could I somehow edit your scripts to build tensorflow/tfjs-node from source?
I came upon this solution
https://github.com/yhwang/node-red-contrib-tf-model
Thanks

[alextran1502]

Yes, please make a PR when you have it working, I will take a look

[alextran1502]

@ippocratis I rework the docker image, it is no longer requires TensorFlow for npm install. I will move TensorFlow to microservice architecture so other people can use different machine learning services if needed. You can reclone the repo and try build on ARM64

[ippocratis]

thanks Dockerfile buils fine now
apreciate it