[SLURM][FSDP][CONTAINER] Docker build fails for slurm example #814
Description
The docker build command in the README.md in Slurm section is docker build -f ../Dockerfile -t fsdp:pytorch2.7.1 . which fails due to incorrect docker context:
=> ERROR [stage-0 5/8] COPY src/ /fsdp/ 0.0s
------
> [stage-0 5/8] COPY src/ /fsdp/:
------
Dockerfile:8
--------------------
6 | RUN ln -s /usr/bin/python3 /usr/bin/python
7 |
8 | >>> COPY src/ /fsdp/
9 | RUN --mount=type=cache,target=/root/.cache/pip pip install -r /fsdp/requirements.txt
10 | RUN pip install hyperpod-elastic-agent
--------------------
ERROR: failed to build: failed to solve: failed to compute cache key: failed to calculate checksum of ref 95e3c277-5b6d-4dfd-88ce-9edbed5e8e1d::1038ja6eurg77w7ujnt5uz6jn: "/src": not found
This can be fixed by changing the docker context. The command needs to be updated to docker build -f ../Dockerfile -t fsdp:pytorch2.7.1 ../. so that the src/ directory comes into the context when building the image.