application: on plugin load failure, log and continue #3486
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Summary: Prior to this change, any exception raised from `loader.load()` would crash TensorBoard (see, e.g., #3484). As of this patch, we log an error, drop the plugin, and continue loading. Test Plan: Run `bazel run //tensorboard -- --logdir /tmp/logs` in a virtualenv with `tensorboard-plugin-wit==1.6.0.post3` but *without* the patch in #3485. Note that TensorBoard now prints: ``` E0406 09:57:12.844471 139872056727360 application.py:260] Failed to invoke plugin loader WhatIfToolPluginLoader.load Traceback (most recent call last): File ".../tensorboard/backend/application.py", line 255, in TensorBoardWSGIApp plugin = loader.load(context) File "/VIRTUAL_ENV/lib/python3.7/site-packages/tensorboard_plugin_wit/wit_plugin_loader.py", line 51, in load version = pkg_resources.parse_version(tensorboard.__version__) AttributeError: module 'tensorboard' has no attribute '__version__' Serving TensorBoard on localhost; to expose to the network, use a proxy or pass --bind_all TensorBoard 2.3.0a0 at http://localhost:6006/ (Press CTRL+C to quit) ``` wchargin-branch: application-load-failure-robustness
|
The |
stephanwlee
approved these changes
Apr 6, 2020
tensorboard/backend/application.py
Outdated
| plugin = loader.load(context) | ||
| except Exception: | ||
| logger.error( | ||
| "Failed to invoke plugin loader %s", |
There was a problem hiding this comment.
Is this supposed to be user visible error string? Not sure how much user cares about "invoking plugin loader". I think it should say something like "Failed to load plugin %s and ignoring it".
There was a problem hiding this comment.
Yes, it’s user-facing, and you’re right, it should be improved. Took
your wording; thanks.
There was a problem hiding this comment.
(It’s slightly awkward because the actual plugin name is given by the
result of load, so all we have is the qualname of the loader itself,
but hopefully “WhatIfToolPluginLoader.load” is clear enough.)
wchargin-branch: application-load-failure-robustness wchargin-source: 368ea5aaa168d2972631207ea3740443cbbf1a43
wchargin-branch: application-load-failure-robustness wchargin-source: 368ea5aaa168d2972631207ea3740443cbbf1a43
stephanwlee
approved these changes
Apr 6, 2020
bileschi
pushed a commit
to bileschi/tensorboard
that referenced
this pull request
Apr 15, 2020
Summary: Prior to this change, any exception raised from `loader.load()` would crash TensorBoard (see, e.g., tensorflow#3484). As of this patch, we log an error, drop the plugin, and continue loading. Test Plan: Run `bazel run //tensorboard -- --logdir /tmp/logs` in a virtualenv with `tensorboard-plugin-wit==1.6.0.post3` but *without* the patch in tensorflow#3485 (i.e., run `git revert e59e9c4`). Note that TensorBoard now prints: ``` E0406 11:00:39.494363 140089282381632 application.py:260] Failed to load plugin WhatIfToolPluginLoader.load; ignoring it. Traceback (most recent call last): File ".../tensorboard/backend/application.py", line 255, in TensorBoardWSGIApp plugin = loader.load(context) File "/VIRTUAL_ENV/lib/python3.7/site-packages/tensorboard_plugin_wit/wit_plugin_loader.py", line 51, in load version = pkg_resources.parse_version(tensorboard.__version__) AttributeError: module 'tensorboard' has no attribute '__version__' Serving TensorBoard on localhost; to expose to the network, use a proxy or pass --bind_all TensorBoard 2.3.0a0 at http://localhost:6006/ (Press CTRL+C to quit) ``` wchargin-branch: application-load-failure-robustness
bileschi
pushed a commit
that referenced
this pull request
Apr 15, 2020
Summary: Prior to this change, any exception raised from `loader.load()` would crash TensorBoard (see, e.g., #3484). As of this patch, we log an error, drop the plugin, and continue loading. Test Plan: Run `bazel run //tensorboard -- --logdir /tmp/logs` in a virtualenv with `tensorboard-plugin-wit==1.6.0.post3` but *without* the patch in #3485 (i.e., run `git revert e59e9c4`). Note that TensorBoard now prints: ``` E0406 11:00:39.494363 140089282381632 application.py:260] Failed to load plugin WhatIfToolPluginLoader.load; ignoring it. Traceback (most recent call last): File ".../tensorboard/backend/application.py", line 255, in TensorBoardWSGIApp plugin = loader.load(context) File "/VIRTUAL_ENV/lib/python3.7/site-packages/tensorboard_plugin_wit/wit_plugin_loader.py", line 51, in load version = pkg_resources.parse_version(tensorboard.__version__) AttributeError: module 'tensorboard' has no attribute '__version__' Serving TensorBoard on localhost; to expose to the network, use a proxy or pass --bind_all TensorBoard 2.3.0a0 at http://localhost:6006/ (Press CTRL+C to quit) ``` wchargin-branch: application-load-failure-robustness
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
Prior to this change, any exception raised from
loader.load()wouldcrash TensorBoard (see, e.g., #3484). As of this patch, we log an error,
drop the plugin, and continue loading.
Test Plan:
Run
bazel run //tensorboard -- --logdir /tmp/logsin a virtualenv withtensorboard-plugin-wit==1.6.0.post3but without the patch in #3485(i.e., run
git revert e59e9c4f376a). Note that TensorBoard now prints:wchargin-branch: application-load-failure-robustness