[CI Problem] Flaky `test_ethosu/test_codegen.py`

[masahi]

This one I'm not sure how frequent it is.

https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-9882/3/
https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-9815/1/

@manupa-arm @leandron @lhutton1 @Mousius

[u99127]

https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-9882/3/

             # process is done if there is no data and the result is valid
               if not data:  # EOF
>                   break
E                   OSError: [Errno 28] No space left on device

We need to fix the CI node - it's run out of space.

[u99127]

https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-9815/1/pipeline

This looks like a catastrophic failure for all the tests for a first look , I'll wait for others who know more about this to comment.

[lhutton1]

https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-9815/1/pipeline

This looks like a catastrophic failure for all the tests for a first look , I'll wait for others who know more about this to comment.

This is #9815 which I believe is expected to be failing at the moment due to needing a CI update. cc @ekalda

[u99127]

Thanks @lhutton1 - so not a flaky test issue I suppose .

[masahi]

Thanks, I missed the No space left on device bits and #9815 was reported by other person as the same flaky failure.

[ekalda]

https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-9815/1/pipeline
This looks like a catastrophic failure for all the tests for a first look , I'll wait for others who know more about this to comment.

This is #9815 which I believe is expected to be failing at the moment due to needing a CI update. cc @ekalda

Yes @lhutton1 is right, this patch is waiting for upstream CI image to be updated so that it would have the U-65 driver :)