Closed elisabettai closed 1 year ago
@mguidon could this be related to the default resource changes?
Also iseg fails super-quick when trying to add a new node.
With iseg, web dev tools gave a 503.
There was also this pydantic error in the director none is not an allowed value (type=type_error.none.not_allowed
)
@mguidon could this be related to the default resource changes?
No. After some debugging with @sanderegg we noticed that the problem was that two services associated to @elisabettai were failing the validation above because all associated containers failed. The service exists but has no container therefore the service_port == None
and therefore the validator RunningDynamicServiceDetails
failed.
This issue needs a follow up: @GitHK we need to determine what is expected from get_running_services
since a service can still exists without any running container.
@GitHK please let me know when you have a moment to check this issue.
This is what happened to @elisabettai :
None
(since they were failed... that makes sense)RunningDynamicServiceDetails
failed validating because it does not accept service_port = None
Workaround: @sanderegg deleted legacy services that had no containers running and everything worked again.
Closing since from user perspective the workaround solved it. Let me know @pcrespov if I should create a follow-up case in osparc-simcore (or if you prefer doing that).
Long Story Short The TVB service fails starting on osparc.io
Steps to reproduce In a new study, add a new tvb node
Additional context This is a legacy dynamic service.
In Portainer,
tvb-app
(backend part of the service) stays pending forever, with error messageno suitable node (insufficient resources on 8 nodes; scheduling constraints not satisfied on 3 nodes)
tvb-web
fails (11 instances are there) withhost not found in upstream "tvb-app_6914e85f-ce9c-4ef3-be7b-b897f1627057:8060" in /etc/nginx/conf.d/default.conf:2
This is what the user sees in the logger:
Service failed: task: non-zero exit (1)