Skip to content

fix: Handle unload too quick after load #5504

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

Merged
merged 5 commits into from
Apr 8, 2024

Conversation

sakoush
Copy link
Contributor

@sakoush sakoush commented Apr 8, 2024

In the cases where we have a model load that is followed by unload/delete too quickly, there is an edge case where the unloading logic is confused because there is no server attached yet (while the model is being loaded from the first cmd). This change then adds another check to make sure that we consider cases where the model is still loading.

fixes INFRA-943 (internal)
Notes:

  • sometimes in the model server agent we get this error now
level=error msg="Failed to handle load model tfsimple1:1" Name=Client error="rpc error: code = Internal desc = State mismatch for tfsimple1:1 expected state Loading but was Unloading when trying to move to state Loaded"

This means that the state has changed beyond the expected flow, which is probably fine in this situation.

@sakoush sakoush requested a review from lc525 as a code owner April 8, 2024 15:40
Copy link
Member

@lc525 lc525 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

lgtm

@sakoush sakoush added the v2 label Apr 8, 2024
@sakoush sakoush merged commit 0b9ee7d into SeldonIO:v2 Apr 8, 2024
jtayl222 pushed a commit to jtayl222/seldon-core that referenced this pull request Jul 20, 2025
* fix note

* add ability to check loading state

* change func name to be more reflective
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants