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.

Sign up for GitHub

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

Jump to bottom

[tune] Raise warning (not an exception) if metadata file not found in checkpoint #33123

Merged
merged 4 commits into from
Mar 8, 2023
Merged

[tune] Raise warning (not an exception) if metadata file not found in checkpoint #33123

merged 4 commits into from
Mar 8, 2023

Conversation

krfricke
Copy link
Contributor

@krfricke krfricke commented Mar 8, 2023

Why are these changes needed?

If a trial is interrupted during checkpointing (e.g. with a timeout, or when firing an interrupt), a half-written checkpoint folder might be left. This checkpoint folder is not cleaned up automatically. If the trial was on the head node, and an ExperimentAnalysis object is subsequently created from the experiment path, the analysis object sometimes will search the local path for existing checkpoints. In that case, it will currently fail because no metadata file has been written.

The correct behavior is to ignore this half-written checkpoint directory. So instead of raising an exception, we should raise a warning instead and ignore the checkpoint.

Related issue number

Checks

  • I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
  • I've run scripts/format.sh to lint the changes in this PR.
  • I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
  • Testing Strategy
    • Unit tests
    • Release tests
    • This PR is not tested :(

Kai Fricke added 2 commits March 7, 2023 18:16
[tune] Raise warning (not an exception) if metadata file not found in…
85f8058 
… checkpoint

Signed-off-by: Kai Fricke <kai@anyscale.com>
continue
0741e7b 
Signed-off-by: Kai Fricke <kai@anyscale.com>
@krfricke krfricke requested a review from Yard1 March 8, 2023 02:19
Kai Fricke added 2 commits March 7, 2023 18:20
Full path
755b250 
Signed-off-by: Kai Fricke <kai@anyscale.com>
nit
027e08d 
Signed-off-by: Kai Fricke <kai@anyscale.com>
@krfricke krfricke merged commit 902cb7e into ray-project:master Mar 8, 2023
@krfricke krfricke deleted the tune/ea-metadata-warning branch March 8, 2023 04:45
ProjectsByJackHe pushed a commit to ProjectsByJackHe/ray that referenced this pull request Mar 21, 2023
@krfricke @ProjectsByJackHe
[tune] Raise warning (not an exception) if metadata file not found in…
fd0e930 
… checkpoint (ray-project#33123)

If a trial is interrupted during checkpointing (e.g. with a timeout, or when firing an interrupt), a half-written checkpoint folder might be left. This checkpoint folder is not cleaned up automatically. If the trial was on the head node, and an ExperimentAnalysis object is subsequently created from the experiment path, the analysis object sometimes will search the local path for existing checkpoints. In that case, it will currently fail because no metadata file has been written.

The correct behavior is to ignore this half-written checkpoint directory. So instead of raising an exception, we should raise a warning instead and ignore the checkpoint.

Signed-off-by: Kai Fricke <kai@anyscale.com>
Signed-off-by: Jack He <jackhe2345@gmail.com>
edoakes pushed a commit to edoakes/ray that referenced this pull request Mar 22, 2023
@krfricke @edoakes
[tune] Raise warning (not an exception) if metadata file not found in…
36c8dad 
… checkpoint (ray-project#33123)

If a trial is interrupted during checkpointing (e.g. with a timeout, or when firing an interrupt), a half-written checkpoint folder might be left. This checkpoint folder is not cleaned up automatically. If the trial was on the head node, and an ExperimentAnalysis object is subsequently created from the experiment path, the analysis object sometimes will search the local path for existing checkpoints. In that case, it will currently fail because no metadata file has been written.

The correct behavior is to ignore this half-written checkpoint directory. So instead of raising an exception, we should raise a warning instead and ignore the checkpoint.

Signed-off-by: Kai Fricke <kai@anyscale.com>
Signed-off-by: Edward Oakes <ed.nmi.oakes@gmail.com>
peytondmurray pushed a commit to peytondmurray/ray that referenced this pull request Mar 22, 2023
@krfricke @peytondmurray
[tune] Raise warning (not an exception) if metadata file not found in…
265e8b1 
… checkpoint (ray-project#33123)

If a trial is interrupted during checkpointing (e.g. with a timeout, or when firing an interrupt), a half-written checkpoint folder might be left. This checkpoint folder is not cleaned up automatically. If the trial was on the head node, and an ExperimentAnalysis object is subsequently created from the experiment path, the analysis object sometimes will search the local path for existing checkpoints. In that case, it will currently fail because no metadata file has been written.

The correct behavior is to ignore this half-written checkpoint directory. So instead of raising an exception, we should raise a warning instead and ignore the checkpoint.

Signed-off-by: Kai Fricke <kai@anyscale.com>
elliottower pushed a commit to elliottower/ray that referenced this pull request Apr 22, 2023
@krfricke @elliottower
[tune] Raise warning (not an exception) if metadata file not found in…
7e43ad6 
… checkpoint (ray-project#33123)

If a trial is interrupted during checkpointing (e.g. with a timeout, or when firing an interrupt), a half-written checkpoint folder might be left. This checkpoint folder is not cleaned up automatically. If the trial was on the head node, and an ExperimentAnalysis object is subsequently created from the experiment path, the analysis object sometimes will search the local path for existing checkpoints. In that case, it will currently fail because no metadata file has been written.

The correct behavior is to ignore this half-written checkpoint directory. So instead of raising an exception, we should raise a warning instead and ignore the checkpoint.

Signed-off-by: Kai Fricke <kai@anyscale.com>
Signed-off-by: elliottower <elliot@elliottower.com>
ProjectsByJackHe pushed a commit to ProjectsByJackHe/ray that referenced this pull request May 4, 2023
@krfricke @ProjectsByJackHe
[tune] Raise warning (not an exception) if metadata file not found in…
efefa1a 
… checkpoint (ray-project#33123)

If a trial is interrupted during checkpointing (e.g. with a timeout, or when firing an interrupt), a half-written checkpoint folder might be left. This checkpoint folder is not cleaned up automatically. If the trial was on the head node, and an ExperimentAnalysis object is subsequently created from the experiment path, the analysis object sometimes will search the local path for existing checkpoints. In that case, it will currently fail because no metadata file has been written.

The correct behavior is to ignore this half-written checkpoint directory. So instead of raising an exception, we should raise a warning instead and ignore the checkpoint.

Signed-off-by: Kai Fricke <kai@anyscale.com>
Signed-off-by: Jack He <jackhe2345@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@Yard1 Yard1 Yard1 approved these changes

Assignees

@Yard1 Yard1

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@krfricke @Yard1