Func deploy python not work.
Hi team.
I used the func deploy --remote to deploy the default Python func. But the Tekton build the source as Java. It looks weird.
Here's my function yaml:
# $schema: https://raw.githubusercontent.com/knative/func/refs/tags/v0.36.0/schema/func_yaml-schema.json
# yaml-language-server: $schema=https://raw.githubusercontent.com/knative/func/refs/tags/v0.36.0/schema/func_yaml-schema.json
specVersion: 0.36.0
name: fn-py-01
runtime: python
registry: <registry>
namespace: ronghuanz40
created: 2025-08-13T15:05:40.187484+08:00
build:
git:
url: <git-url>
builder: pack
run:
envs:
- name: name
value: sahara6
deploy:
serviceAccountName: ronghuanz40
options:
resources:
requests:
cpu: 50m
memory: 129Mi
limits:
cpu: 129m
memory: 1024Mi
scale:
min: 0
max: 3
kubectl logs --tail=100 -f fn-py-01-00001-deployment-86f8c8dd75-lv2cw -n ronghuanz40
Calculated JVM Memory Configuration: -XX:MaxDirectMemorySize=10M -Xmx479908K -XX:MaxMetaspaceSize=56667K -XX:ReservedCodeCacheSize=240M -Xss1M (Total Memory: 1G, Thread Count: 250, Loaded Class Count: 7591, Headroom: 0%)
Enabling Java Native Memory Tracking
Using readonly truststore: /tmp/truststore
Adding 146 container CA certificates to JVM truststore
---> Installing application source ...
mv: cannot stat '/tmp/src/*': No such file or directory
func version -v
Version: v0.45.1
Knative: v1.18.1
Commit: c1160e14
SocatImage: ghcr.io/knative/func-utils:v2
TarImage: ghcr.io/knative/func-utils:v2
Hey! thank you very much for reporting this issue. Im able to reproduce.
So far it seems that the issue is only with python. The other runtimes seem to work as intended. I suspect there is an issue because of our recent python template re-do. Will investigate further.
Hey! thank you very much for reporting this issue. Im able to reproduce.
So far it seems that the issue is only with python. The other runtimes seem to work as intended. I suspect there is an issue because of our recent python template re-do. Will investigate further.
Thanks for your reply. Kindly let me know if there are any updates. Thanks.
@gauron99 Hi, I just wanted to check if there are any updates on this issue. Am I able to use the Python function now? Thanks a lot.
@ryan-alexander-zhang hey, Im pretty sure I have identified the issue, its with the buildpacks and the way we package the function. Im now working on a fix. Theres an "easy" solution but that was a bit too hacky to merge to main so its taking a bit more time. Thanks for your patience! PR should be up soon. I expect couple of days max. This week for sure.
Thanks so much for your reply and for all the hard work. It's really appreciated.
@ryan-alexander-zhang I have a working change. If you were interested in doing a bit of hacking you can try already using my branch, otherwise I will get this into func after some time. Im just also doing a bit of related cleanup etc.
If you were to use my branch https://github.com/gauron99/func/tree/push-loznlwuozlqs you would ~~need to build the func-util image and~~ use that instead of the func one (you can see a where Im going within the changes in the branch).
edit: Actually since I already built the image, you might not need to do much hacking after all, all changes should be within the branch so just make and run it.
Let me know if you are interested in that and I can get you the necessary steps, otherwise Ill get it into func asap.
@gauron99 Thanks—that’s really helpful. I read through the HEAD commit in your repo, and it looks like I just need to rebuild the CLI using the image you’ve already built.
And I have a question about pkg/pipelines/tekton/pipelines_provider.go: is this where the Tekton scaffolding step is implemented?
The scaffolding step is really in a transition right now. Think of this as a hot fix for an actual codebase fix later. I have that planned right after this. And yes, the scaffolding is done vie the "injector" in the new changes in my branch
and yes, you should be able to just rebuild!
Thank you very much! I will have a try.