add workload runtime support for aks #3896
Conversation
alexeldeib
requested review from
andyliuliming and
zqingqing1
as code owners
alexeldeib
force-pushed
the
ace/wasm
branch
from
086a8bd to
b9c4659
Compare
|
aks |
|
seems like 429? will retry tomorrow :/ |
| text: az aks nodepool add -g MyResourceGroup -n nodepool1 --cluster-name MyManagedCluster --os-sku Ubuntu | ||
| - name: Create a nodepool which can run wasm workloads. | ||
| text: az aks nodepool add -g MyResourceGroup -n nodepool1 --cluster-name MyManagedCluster --workload-runtim WasmWasi |
| @@ -141,6 +142,7 @@ def load_arguments(self, _): | |||
| c.argument('assign_kubelet_identity', type=str, validator=validate_assign_kubelet_identity) | |||
| c.argument('disable_local_accounts', action='store_true') | |||
| c.argument('yes', options_list=['--yes', '-y'], help='Do not prompt for confirmation.', action='store_true') | |||
| c.argument('workload_runtime', arg_type=get_enum_type([CONST_WORKLOAD_RUNTIME_OCI_CONTAINER, CONST_WORKLOAD_RUNTIME_WASM_WASI])) | |||
There was a problem hiding this comment.
I suppose it would be better if you could create a global variable such as workload_runtimes = [CONST_WORKLOAD_RUNTIME_OCI_CONTAINER, CONST_WORKLOAD_RUNTIME_WASM_WASI], then set this variable to get_enum_type, which would avoid forgetting to modify the optional parameters of create or nodepool add in the future.
There was a problem hiding this comment.
From help message, I suppose we should add default=CONST_WORKLOAD_RUNTIME_OCI_CONTAINER?
alexeldeib
force-pushed
the
ace/wasm
branch
from
17a1727 to
58d0f22
Compare
|
@FumingZhang waiting for anything here besides re-running live tests? |
|
@alexeldeib Please write those changes into the By the way, do you need to release a new version for these changes? If so, please also update the |
|
In addition, could you please address the live test issue? |
alexeldeib
force-pushed
the
ace/wasm
branch
from
58d0f22 to
3611889
Compare
yes please :) bumped to 0.5.34
@zhoxing-ms these don't seem related, since the newly added test is a preview feature and doesn't run in the live pipeline. the previous failures I saw seemed throttling related (ARM 429). the new commits triggered a fresh build, hopefully that works. in the future, is there a good way to manually force a rerun? I did not see one. |
|
@zhoxing-ms @hbeberman looks like the failing test is related to mariner, I suspect due to a known issue with extensions @hbeberman discovered this week. I'm not sure whether we should disable the live mariner test here, or wait for the RP fix to unblock existing PRs? |
|
Yes, I can help confirm that the failed test case is due to compatibility issues with CBLMariner and has nothing to do with the changes in this PR. So I suppose we could bypass the check and merge the PR for this time. I am working on a PR #3927 to exclude cases like this. |
This checklist is used to make sure that common guidelines for a pull request are followed.
General Guidelines
azdev style <YOUR_EXT>locally? (pip install azdevrequired)python scripts/ci/test_index.py -qlocally?For new extensions:
About Extension Publish
There is a pipeline to automatically build, upload and publish extension wheels.
Once your PR is merged into master branch, a new PR will be created to update
src/index.jsonautomatically.The precondition is to put your code inside this repo and upgrade the version in the PR but do not modify
src/index.json.