Closed matthewcarbone closed 1 year ago
To be clear, this will push a new image to the NERSC container registry. It won't change anything about the pods or anything else in the running cluster.
There probably are limits to the number of images we can have, but publish-on-merge is a standard pattern and I would expect the limits to be high enough to enable that. If we hit a limit, one recourse available is to cull old images, retaining only tagged images and the most recent main
/dev
images.
This change worked as expected, the workflow got further, but still failed on
#3 [internal] load metadata for ghcr.io/bluesky/tiled:0.1.0a74
#3 ERROR: ghcr.io/bluesky/tiled:0.1.0a74: not found
------
> [internal] load metadata for ghcr.io/bluesky/tiled:0.1.0a74:
------
ERROR: failed to solve: rpc error: code = Unknown desc = failed to solve with frontend dockerfile.v0: failed to create LLB definition: ghcr.io/bluesky/tiled:0.1.0a74: not found
Error: buildx failed with: ERROR: failed to solve: rpc error: code = Unknown desc = failed to solve with frontend dockerfile.v0: failed to create LLB definition: ghcr.io/bluesky/tiled:0.1.0a74: not found
I'm not sure how to interpret that error. The image is there: https://github.com/bluesky/tiled/pkgs/container/tiled/41564249?tag=v0.1.0a74
The issue is that the tag is v0.1.0a74
not 0.1.0a74
. See #59.
@danielballan are we sure we want a new version being pushed on every merge to
main
/dev
? That'll be ok server-side? I don't expect we'll push changes that often but I just want to be sure that there's no "limit" or something on how many pods we can have.