[FLUME-3341] Taildir source may cause file handle leak and data duplication

• Type: Bug
• Status: Patch Available
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.8.0, 1.9.0
• Fix Version/s: 1.8.0
• Component/s: Sinks+Sources
• Release Note: Flume 1.8.0 is fixed
• Flags: Patch
• Assignee: Unassigned
• Reporter: Ming Zhekai
• Created: 18/Aug/19 14:05
• Updated: 20/Aug/19 02:29

Description

As is described in Flume-3342, renaming a file may cause data duplication. But moreover, if the original file was opened by flume before and not closed yet, flume reopens it while not freeing its original file handles. And then I delete the new file, but after the idle time, flume just closes the new file handle and forgets to close its original file handles, leading to a handle leak. It can be found by "lsof | grep "deleted".

This bug is triggered by using  Log4j to roll log files. To avoid losing data when rolling, I use regex to include both current log file and old log files in the taildir path.

 

Issue Links

• Blocked: FLUME-3216 When I rename the file, it causes flume to read the data repeatedly.
• Blocked: FLUME-3219 Taildir source: if file is renamed, it is consumed again
• links to: GitHub Pull Request #295