-
Notifications
You must be signed in to change notification settings - Fork 9k
HDFS-17707. TestStateStoreFileSystem fails after HDFS-17496. #7276
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
Conversation
@Hexiaoqiao @kokon191 Sir, PTAL when you have free time, thanks a lot. |
💔 -1 overall
This message was automatically generated. |
@hfutatzhanghb Sorry, I plan to roll back #6764 until the fix is complete before merging. Looking forward to a smooth repair, thank you again! cc: @Hexiaoqiao |
@slfan1989 Got it sir. Thanks a lot for your operating. Will fix it soonly. |
I'm leaning towards the 1st approach more. 1 lock per directory is more intuitive. |
8b99828
to
2a418dc
Compare
💔 -1 overall
This message was automatically generated. |
Hi @hfutatzhanghb is this PR still waiting to review considering #6764 was revert now. Please close it if not yet. Thanks. |
Description of PR
JIRA: HDFS-17707. TestStateStoreFileSystem fails after HDFS-17496.(#6764)
Cause is race condition when multiple instances of FileIoProvider#mkdirsWithExistsCheck are called for the same directory on different DIR locks.
Actually, ModDataSetSubLockStrategy is thread-safe.
There can be two ways to solve this problem, I prefer to the first approach:
File#mkdirs() will return false when dir is already existed.