shanrui opened a new pull request, #8542:
URL: https://github.com/apache/hadoop/pull/8542
Problem:
In modern libfuse implementations, fi->flags in flush() callback
may be reset to 0. The original fuse_dfs code relies on:
if (fi->flags & O_WRONLY)
As a result, hdfsHFlush/hdfsHSync is skipped, causing newly written
data to be invisible to readers until file close.
Root cause:
fi->flags is not guaranteed to be preserved in flush().
Fix:
Store open flags in dfs_fh during dfs_open(), and use it in flush()
instead of fi->flags.
Testing:
- Verified write -> flush -> read consistency
- Verified hdfsHFlush is triggered correctly
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
