I am writing my own plugin where I need to set the public URL in an application config file using an Ansible task.
In the logs of a workspace I see a service_url
key with value https://d08d27c8-d63d-492f-9479-70aa98b90f81.workspaces.live.surfresearchcloud.nl
.
When I try to use that var in my Ansible task, then it is not set.
Can this service_url
var be supplied to where my plugin is being run?
I am writing my own plugin where I have 2 storage items (mainly because system disk is too small):
- A read only disk with data
- A scratch disk with intermediate results for all CO members I looked at the resource_meta dict, but it only contains internal uuids, not the name I gave my storage items. In a Ansible task how do I distinquish between them as they need to be mounted differently?
I am starting workspace with my own application+plugin. When things fail the logs show a long line with the output of my plugin. However when my plugin completes succesfully I don't see any mention of my plugin tasks in the log. I am printing some messages in the plugin about where parts of the application are running, but I can not see them in the logs. Can the output of external plugins always be logged just like the internal plugins (RSC-OS, RSC-CO, RSC-Nginx)?
For my plugin I need to install conda. There are some ready made roles on Ansible Galaxy like https://galaxy.ansible.com/evandam/conda that I would like to use.
To use locally I would create requirements.yml
for example https://github.com/eWaterCycle/infra/blob/5efee1c5915aa4a8157c50053b093557944ccd5c/requirements.yml . And then run ansible-galaxy install -r requirements.yml
to install those roles. Can the RSC-External plugin be extended to install the roles from Ansible Galaxy?
Stuff to ask Research Cloud helpdesk
- How to become root after creation, see https://servicedesk.surfsara.nl/wiki/display/WIKI/RC+plugin+passwordlesssudo
- How to add users to docker posix group
- How to mark certain user as JupyterHub admin -> use sram api to sync posix group
- How to put home dirs on attached storage
- What is symlink to /scratch doing in homedir? How to tie storage item to it?
- each user has symlink ~/data -> /data, which has disks as subdirs
- each user has symlink ~/scratch -> /scratch , which points to root disk
- Fetch workspace.json from https//gw.live.surfresearchcloud.nl/workspace/workspaces gives "Access to this API has been disallowed" error
- Show set chosen parameters in running workspace, now don't know what I picked