[BUG] terminate called after throwing an instance of 'PAL_SEHException' on Mac silicon chip M1 M2

[yzlee]

Describe the bug

If you are using mac silicon and vscode dev container. You will meet these error:

ERROR: rosetta error: failed to open elf at /lib64/ld-linux-x86-64.so.2

Assigning 'Cognitive Services OpenAI Contributor' role to managed identity... ________________________________
/ Uh oh, an error has occurred.
\ Please see message below. /

and

Deployment name: graphrag-deploy-2024-08-31_23-19-14
ERROR: terminate called after throwing an instance of 'PAL_SEHException'
Assigning 'Cognitive Services OpenAI Contributor' role to managed identity...  ________________________________
/  Uh oh, an error has occurred. \
\  Please see message below.     /
 ‾‾‾‾‾‾‾‾‾‾/‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
          /
      __ /
     /  \
    ~    ~
   / \  /_\
   \o/  \o/
    |    |
    ||   |/
    ||   ||
    ||   ||
    | \_/ |
    \     /
     \___/

Unable to parse service principal id from azure outputs, exiting...
graphrag-solution-accelerator-py3.10

It's not your fault. what you need do is just modify Dockerfile:

1. change the first line to FROM --platform=linux/amd64 python:3.10
2. add the second line ENV DOTNET_EnableWriteXorExecute=0

below is the principle explaination: https://stackoverflow.com/questions/77105077/how-to-run-docker-in-devcontainers-on-mac-m1-apple-silicon https://stackoverflow.com/questions/78578636/how-to-get-dotnet-8-azure-function-to-run-on-an-m3-chip-inside-a-docker-containe

To Reproduce i am the owner of the resource group. I am the owner of subscription. Everything is set following https://github.com/Azure-Samples/graphrag-accelerator/blob/main/docs/DEPLOYMENT-GUIDE.md I am using macbook M2.

Expected behavior echo is misleading

Desktop (please complete the following information):

• OS: MacOS
• Version 14.6.1
• M2 chip

[gmpark-factagora]

I had the same problem and the proposed solution worked. Thank you for the information!

[vlad-tsoy]

This has also worked for me. This TIP should be added to the FAQ IMHO. Thank you!