Expose the workstation’s local file system to Docker from DSS

RajeshRana84

Good Day All,

I am looking to use Docker as containerized environment and for the same i wanted to use Docker bind-mount. I am not sure how to expose the workstation’s local file system to Docker from DSS.

Docker allows such functionality using bind-mount https://docs.docker.com/storage/bind-mounts/

So what i am looking for is something like this :

docker run -d -it --name test-container --mount type=bind,source=/tmp,target=/usr nginx:latest

How can we pass these parameters (--mount type=bind,source=/tmp,target=/usr) from Dataiku Containerized Execution page ?

Sergey

Hi @RajeshRana84

It is not possible to mount additional volumes into docker containers spun-up for the containerized execution.

Can you please share what is the use case doing this?

RajeshRana84

Hi @sergeyd
,

Thanks for the response. I am looking to have a persistent storage to store data on my local system. Is it possible ?

Thanks,
Rajesh

Sergey

Hi @RajeshRana84

In case you need to have the same storage available for local and containerized execution, you may want to check on managed folders:

https://doc.dataiku.com/dss/latest/connecting/managed_folders.html

You can mount a volume to DSS, create a managed folder based on it and access it from the containers via Python or R APIs.

RajeshRana84

Good Day @sergeyd
,

Thanks for the information. But for me the scenario is a bit different.

So i have a server A where DSS is running and then there is server B where docker is running.

Now DSS (on server A) uses the docker execution environment on server B. And there is very large file say 3 TB which docker need as input for further execution. Now if i use managed filesystem then i have to place this file on server A and then transfer the file to server B , which is not feasible.

Now in the above scenario , is there a way i can use the 3 TB file on Server B so that Docker can take it as input and proceed.