Examples: query, "exact match", wildcard*, wild?ard, wild*rd
Fuzzy search: cake~ (finds cakes, bake)
Term boost: "red velvet"^4, chocolate^2
Field grouping: tags:(+work -"fun-stuff")
Escaping: Escape characters +-&|!(){}[]^"~*?:\ with \, e.g. \+
Range search: properties.timestamp:[1587729413488 TO *] (inclusive), properties.title:{A TO Z}(excluding A and Z)
Combinations: chocolate AND vanilla, chocolate OR vanilla, (chocolate OR vanilla) NOT "vanilla pudding"
Field search: properties.title:"The Title" AND text
Answered
I’M

I’m working on an automated deployment of ClearML with IaC. I’ve got a script to start an EC2 instance that runs the docker compose file.

Separately, I’ve got an aws_autoscaler.yaml file that correctly configures aws_autoscaler.py with the right IAM permissions, startup commands, etc.

The missing piece I need to automate is: I need to add a single worker to the services queue when the EC2 instance starts up so that I can run the autoscaler task. I’ll run that worker on the same EC2 instance since the services queue should only handle light tasks.

But is there any way I can auto-generate the credentials needed for that first services worker/agent instance? Or is there a way I could get around creating them for that particular worker?

  
  
Posted 2 years ago
Votes Newest

Answers 10


That could work! Is that an option? Something that lets me spin up the ClearML and get a services worker to connect to it without manual steps.

  
  
Posted 2 years ago

I did a quick local experiment and observed that credentials created from the UI indeed become invalid if you delete the ClearML volumes.

  • starting docker-compose locally
  • creating a set of credentials from the UI
  • hardcodign those credentials into the docker-compose file
  • restarting
  • the agent-services container started up and successfully became a registered worker
  • I killed the docker-compose and deleted the volume folders
  • restarted the docker-compose (with the same hard-coded credentials)
    The resulting logs from the agent-services container were this:
Using environment access key CLEARML_API_ACCESS_KEY=NL0V1RKGB0TN57L3OZRK
Using environment secret key CLEARML_API_SECRET_KEY=********

clearml_agent: ERROR: Failed getting token (error 401 from 
): Unauthorized (invalid credentials) (failed to locate provided credentials)
  
  
Posted 2 years ago

yes 🙂

  
  
Posted 2 years ago

I SOLVED IT, NO NEED TO READ FURTHER 😄

I'm a chump and didn't read the docs: None


Oh, I think I got overexcited and didn't look at this closely. So this ACCESS/SECRET key pair is on the agent-services container.

I can see that agent-services is simply a container running clearml-agent daemon --queue services --docker
and hey! that's what I was going to have to add to achieve getting a services worker!

I do have a question: Is there a way I can get the CLEARML_API_ACCESS_KEY and CLEARML_API_SECRET_KEY programmatically to give to this service? Or do I need to create these in the UI? If I create them in the UI and then destroy/re-create the EC2 instance the docker-compose.yml is running on, would that invalidate those keys?
image

  
  
Posted 2 years ago

One idea: is it possible to store usable credentials in advance and place them in a volume that the ClearML containers can access and then use?

  
  
Posted 2 years ago

This is already part of the docker-compose file,
https://github.com/allegroai/clearml-server/blob/master/docker/docker-compose.yml

  
  
Posted 2 years ago

^^^ For my own notes: this is the web request made by the frontend to create a set of credentials
image

  
  
Posted 2 years ago

The question I'm exploring remains: is it possible to acquire that initial set of ClearML API keys programmatically so that the manual steps of 1-4 above can be avoided for an initial deployment?

  
  
Posted 2 years ago

Oh my word, is this it? None

Can you set these to any strings and have them function as API keys?

  
  
Posted 2 years ago

@<1541954607595393024:profile|BattyCrocodile47> you mean like environment variables?

  
  
Posted 2 years ago
2K Views
10 Answers
2 years ago
2 years ago
Tags
Similar posts