I set ClearML up with Kubernetes Helm.

Got it!

Since k8s has no real priority/order the scheduling order is not guaranteed form this point.

K8s can schedule pod with different priorities. So maybe no real scheduling means there is no ClearML scheduling after applying pod to k8s.

only if it is sure there are enough resources to actually spin the job now

In 1.0.2 version, I do not find any logic about checking k8s resource in k8s.py. Does it will implement in the future?

I want to know the total fields(string reference mentioned above) which can be used as input of step2. Are these fields of ClearML Task?

Another question: can I mix steps with Task and Function?

According to pipeline_from_functions.py , it is easy to understand that step1 returns data_frame and I can use it as input of step2. But I have no idea what string reference could be used when steps come from Task?

code

pipe.add_function_step(
    name='step_one',
    function=step_one,
    function_kwargs=dict(pickle_data_url='${pipeline.url}'),
    function_return=['data_frame'],
    cache_executed_step=True,
)
pipe.add_function_step(
    name='step_two',
    # parents=['st...

Hi, TimelyPenguin76 , could you help me?

For example, I have a lot of tasks in a queue, but there is a default agent(services, mode = daemon) after using helm to deploy the ClearML in Kubernetes. As far as I know, agent will pull and execute the task in specified queue one by one, which results in other tasks are blocked. However, the physical resource is enough in cluster. So how can ClearML to improve the efficiency of execution? I can solve it by implementing a custom scheduler which is used to watch the queue, pull the task and...

But you’re not using the webserver, you’re sending the cURL requests manually, aren’t you?

yes, I send them manully.

In any case, the said credentials are valid for the company ID defined using 

apiserver.default_company

It really works for apiserver, but webserver does not have the default_company configuration.

Do you use ClearML Open Source?

Hi, SuccessfulKoala55 Maybe this is not a good question. In other words, docs introduce that ClearML Open Source supports orchestration, how can I found the relating codes? And what the role of clearml-agent in orchestration, a combination of kube-scheduler and kubelet?

Thanks for your reply 😀 ！

I have another question. Does the following apis in the picture are served by both of apiserver and webserver, or just one of them ?

And no real scheduling in README exactly means that ClearML will not do scheduling for task which may provided in paid tier? In other words, Kubernetes will schedule the Pod for me. Is it right?

Also - did you start with a fresh DB after changing the default company value?

No, I just restart the apiserver and webserver pod.

Thanks a lot! I’ ll replace credentials with new user in assigned to a new company ID to call users.get_all again.

How can I fresh DB? Mongo or Redis?

Hi BurlySeagull48 AgitatedDove14 , does upgrading clearml package and images to 1.0.4 really work? https://clearml.slack.com/archives/CTK20V944/p1628112607184300?thread_ts=1628108536.180100&cid=CTK20V944

AgitatedDove14 , I will try it with latest version. But why the url in es is different from it in web UI? I have checked the codes that evnets.debug_images api which will return url to web UI will get the url from es.

code commit

I set it up with the following documentation: https://allegro.ai/clearml/docs/docs/deploying_clearml/clearml_server_kubernetes_helm.html

Hi  AgitatedDove14 ,I try it with Clearml 1.1.4, but it fails again.

README introduces that no real scheduling when using Kubernetes. But I think no matter launch a Pod or Job that the Pod will be scheduled through k8s scheduler exception specifying the nodeName in Pod specification. This confuses me a lot. What is the meaning of no real scheduling I don’t understand no verification of docker image

What is meaning of ‘the scheduling is done before that’? Does it mean the ClearML will schedule the task for me?

Thanks