[CASSANDRA-9482] SSTable leak after stress and repair

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Duplicate
• Labels: stress
• Severity: Normal
• Assignee: Marcus Eriksson
• Reporter: Jim Witschey

Authors

Marcus Eriksson

• Created: 26/May/15 17:28
• Updated: 16/Apr/19 09:31
• Resolved: 29/May/15 08:02

Description

I have a dtest that fails intermittently because of SSTable leaks. The test logic leading to the error is:

• create a 5-node cluster
• insert 5000 records with stress, RF=3 at CL=ONE
• run flush on all nodes
• run repair on a single node.

The leak is detected on a different node than repair was run on.

The failing test is here. The relevant error his here, along with the errors from the entire 5-node cluster. In these logs, the repair was run on node1 and the leak was found on node2.

I can bisect, but I thought I'd get the ball rolling in case someone knows where to look.

Issue Links

• duplicates: CASSANDRA-9508 Avoid early opening of compaction results when doing anticompaction