Agent works when I am running it from virtual environment but stucks in the same place all the time when I using Docker
Can you please provide a log? I'm not sure what it means stuck
No 🙂
The full process to get the same code you run on your machine includes applying the local changes:
Clone the code from a git repository. Apply any uncommitted changes recorded.you can view the entire clearml agent diagram https://clear.ml/docs/latest/docs/clearml_agent
AgitatedDove14 do I need to have the repo that I am running on my account? Even if it is public repo, like repo with your (clearml) examples:
SOURCE CODE
REPOSITORY
https://github.com/allegroai/clearml.git
BRANCH NAME
Latest in branch master
SCRIPT PATH
pytorch_matplotlib.py
WORKING DIRECTORY
examples/frameworks/pytorch
?
There is a git repo 🙂 my question was to clarify if I understand well. Thank you for response :)
AgitatedDove14 how does the Agent know which git repo from my account to clone for execution?
So there is no way to use Agent without use of remote repo (just using local server not connected to Internet), am I right?
Hi RoundMosquito25
How did you spin the agent (whats the cmd line? is it in docker mode or venv mode?)
From the console it seems the pip installation inside the container (based on the log this is what I assume) seems like it is stuck ?!
btw. why do I need to give my git name/pass to run it if I serve an agent from local?
(just using local server not connected to Internet), am I right?
You can if you host your own git server, Or if your code is a single file / jupyter notebook, then the entire code is stored on the Task.
btw: what is the exact setup, how come there is no git repo?
do I need to have the repo that I am running on my account
If it is a public repo, then no need, credentials are only needed for private repos 🙂
Am I missing something ?
Its stored on the Task, you can see it under the execution tab in the UI
btw. why do I need to give my git name/pass to run it if I serve an agent from local?
The main idea is that you can run the agent in any machine (local, cloud) and all should be done out of the box.
If your code is running as part of a git repository, the clearml agent will have to clone it, and for doing so it will use credentials.
Git name and pass are one way for it, but you can also use ssh - if you dont have the git name and password in the configuration, the clearml-agent will try cloning the repo with ssh (it should be printed in the log too), you can also enforce it with agent.force_git_ssh_protocol: true in your clearml.conf file.
Ok, I noticed something that might have been causing that. I didn't add "agent" section to config file...
Agent works when I am running it from virtual environment but stucks in the same place all the time when I using Docker
RoundMosquito25 do notice the agent is pulling the code from the remote repo, so you do need to push the local commits, but the uncommitted changes clearml will do for you. Make sense?
Do I need to push the needed code to github if it needs to be cloned?
Actually I am still struggling with a problem of agent running on docker (message on starting at 10:54)