Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Inspect fails if model was undeployed before getting it info #507

Closed
ablatov opened this issue Oct 3, 2018 · 3 comments
Closed

Inspect fails if model was undeployed before getting it info #507

ablatov opened this issue Oct 3, 2018 · 3 comments
Assignees
Labels
bug [Fixed] for any bug fixes.

Comments

@ablatov
Copy link
Member

ablatov commented Oct 3, 2018

Preconditions.
3 test models:

  • id : "demo-abc-model", version : 1.0
  • id : "demo-abc-model", version : 1.1
  • id : "edi-test-model", version : 1.2
    Steps to reproduce:
  1. Run tests (deploy,undeploy,scale,inspect operations) in parallel from
    3.0_check_edi_one_model.robot SUITE
    (model=> id : "edi-test-model", version : 1.2 is used)
    and from
    4.0_check_edi_multi_versioned_models.robot SUITE
    (models=> id : "demo-abc-model", version : 1.0 and version : 1.1 are used).

Expected result

  1. All tests passed because they use different models and do not affect each other.

Actual result

  1. Some tests (minimum 1 from the first and 1 from the second) fail because EDI server inspect command get list of models first and then try to get their info in a loop. While this happens some models have been already destroyed and inspect could not return info and failed with the exception.
    This was tried to be fixed in Inspect fails if cannot find deployment #438 but maybe it is something new.

Possible solutions
Make inspect EDI command more simple or multi thread oriented.
Possible problem method:
https://github.com/legion-platform/legion/blob/develop/legion/legion/edi/server.py#L246

@ablatov ablatov added the bug [Fixed] for any bug fixes. label Oct 3, 2018
@ablatov
Copy link
Member Author

ablatov commented Oct 3, 2018

EDI_error_log.txt

@ablatov
Copy link
Member Author

ablatov commented Oct 3, 2018

ablatov added a commit that referenced this issue Oct 17, 2018
* [#133] Add first patch (without tests edition)

* [#133] add the third test bare model

* [#133] change new model id + fix jenkins file

* [#133] include pabot in jenkins file + refactor jenkins tests

* [#133] restructure tests for the pabot

* [#133] jenkins model tests execution order fix

* [#133] init cookies only once

* [#133] path to profiles changes

* [#133] update get cookies logic for parallel tests

* [#133] fix jenkins_dex_client

* [#133] wait jenkins_dex_client end run

* [#133] fix pipeline and getting cookies

* [#133] fix cookies global vars

* [#133] add logging

* [#133] add more logging

* [#133] fix jenkins tests

* [#133] remove extra logging

* [#133] Rearrange tests

* [#133] add copyrights to test-bare-models, update flask version

* [#133] add fluent_id tests to the common suite

* [#133] fix paths to resources and pipeline

* [#133] fix cookies handling + airflow test

* [#133] fix model version for single model edi tests

* [#133] fix model id for single model edi tests

* [#133] add teardown to the last test for single model edi suite

* [#133] remove extra spaces

* [#133] deploy single edi suite to the second enclave

* [#507] Fixed legion k8s module, removed exception on unavailable model info

* [#507] Fixed legion k8s module, reload deployment data where needed

* [#133] remove redurant enclave and reloading data for the ModelService

* [#133] Minor code refactoring

* [#507] Add catchable exception for non-existent deployment
@ablatov
Copy link
Member Author

ablatov commented Oct 24, 2018

fixed in #461

@ablatov ablatov closed this as completed Oct 24, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug [Fixed] for any bug fixes.
Projects
None yet
Development

No branches or pull requests

2 participants