buf_file: Skip and delete broken file chunks during resume. fix #1760 #1874
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
repeatedly
commented
Feb 28, 2018
|
|
||
| def handle_broken_files(path, mode, e) | ||
| log.error "found broken chunk file during resume. Deleted corresponding files:", :path => path, :mode => mode, :err_msg => e.message | ||
| # After support 'backup_dir' feature, these files are moved to backup_dir instead of unlink. |
Member
|
Looks good. If you push tests in this PR. I'll also see it. |
Member
|
minor comment: It seems better to prepare a method for |
Member
Author
|
Add tests |
daipom
pushed a commit
that referenced
this pull request
Jun 27, 2025
**Which issue(s) this PR fixes**: * Related to #3970 **What this PR does / why we need it**: This PR improves meta file corruption checking. The meta file contains at least the following field values. https://github.com/fluent/fluentd/blob/fa2eb58922e1c36f83bf1d5243b325a860f72864/lib/fluent/plugin/buffer/file_chunk.rb#L249-L254 This PR reinforces #1874. Without this changes, it might causes following error when launch fluentd every time with broken meta file: ``` 2025-06-06 12:11:26 +0900 [error]: unexpected error while checking flushed chunks. ignored. error_class=NoMethodError error="undefined method '<' for nil" 2025-06-06 12:11:26 +0900 [error]: /Users/watson/src/fluentd/lib/fluent/plugin/output.rb:1479:in 'block in Fluent::Plugin::Output#enqueue_thread_run' 2025-06-06 12:11:26 +0900 [error]: /Users/watson/src/fluentd/lib/fluent/plugin/buffer.rb:548:in 'block in Fluent::Plugin::Buffer#enqueue_all' 2025-06-06 12:11:26 +0900 [error]: /Users/watson/src/fluentd/lib/fluent/plugin/buffer.rb:542:in 'Array#each' 2025-06-06 12:11:26 +0900 [error]: /Users/watson/src/fluentd/lib/fluent/plugin/buffer.rb:542:in 'Fluent::Plugin::Buffer#enqueue_all' 2025-06-06 12:11:26 +0900 [error]: /Users/watson/src/fluentd/lib/fluent/plugin/output.rb:1479:in 'Fluent::Plugin::Output#enqueue_thread_run' 2025-06-06 12:11:26 +0900 [error]: /Users/watson/src/fluentd/lib/fluent/plugin_helper/thread.rb:78:in 'block in Fluent::PluginHelper::Thread#thread_create' ``` If the timekey value is corrupted, the above error occurs. Since there is no appropriate way to check timekey directly, check `id`, `c`, and `m` fields instead. This is because when timekey is broken, other fields may also be broken. It might be possible that the `@size` is 0. `@unique_id`, `@created_at`, and `@modified_at` are set when FileChunk is initialized, so they definitely have some values. I think these fields should be written in meta file. So, this PR adds the `id`, `c`, and `m` fields check. Previously, it operates using default value if metadata was broken. However, it can miss the corruption and result in unexpected errors. So, this PR enhances the detection of broken metadata files instead of using defalut value. This change has backward compatible with v0.14 behavior. **Docs Changes**: Not necessarily required. **Release Note**: buf_file: reinforce buffer file corruption check --------- Signed-off-by: Shizuo Fujita <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
On k8s environment, we received the report "file chunks are unexpectedly broken and it causes error loop during resume."
This patch mitigates this problem by ignoring broken chunks in resume.