
Reputation
Badges 1
53 × Eureka!this means network issues at some level
but I think this behaviour will hange in future releases
I finally found where the issue is, I opened an issue on gh so it's more manageable:
https://github.com/allegroai/clearml/issues/273
I also found a not so good (at least for me) behaviour:
https://github.com/allegroai/clearml/issues/272
SuccessfulKoala55 after looking at the issue Iβm a bit confused π ; as far as I can see there is no way to pass any parameter to clearml-agent in daemon mode to push log to stdout. Can you confirm it? (If yes I need to find some workaround)
btw I think it's related clearml-task
usage; did you try to create a pipeline with decorators creating it with clearml-task
before?
k8s cluster can access ubuntu archive?
ReassuredArcticwolf33 pls try this one and let me know if this is working for you
If I'm not wrong you can simply label the namespace to avoid istio to get there
so do you want to mount files into agent pod?
pretty strange, I also noticed in example from line 2from clearml import TaskTypes
it will be easier for me to track fixes
Also k8s distribution and version you are using can be useful
with that in place k8s should be able to provisione pvc
If you have ALB you will just need to add some annotations on ingress rules depending on your setup. btw for now , since you already have everything in place, I suggest to just add values to /etc/hosts
and see if it works
Hi PleasantGiraffe85 , just to get some more info; what version of chart are you using? Did you enabled ingress?
Iβm just trying to understand of itβs something related ceph or clearml deployment
ya sure, I was referring to. create a new PVC just for the test
how you cluster reacts is about scaling infra as much as needed (karpenter or any other cloud autoscaler should work)