[LUCENE-9661] Another classloader deadlock?

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 8.0, master (9.0)
• Fix Version/s: 8.x, master (9.0), 8.8
• Lucene Fields: New
• Assignee: Namgyu Kim
• Reporter: Michael McCandless
• Created: 09/Jan/21 14:31
• Updated: 20/Jan/21 15:59
• Resolved: 20/Jan/21 15:57

Description

The java processes spawned by our Lucene nightly benchmarks sometimes randomly hang, apparently while loading classes across threads, under contention.

I've opened this luceneutil issue with some details, but Uwe Schindler suggested I open an issue here too since he has been seeing this in CI builds too.

It is rare, maybe once a week in the nightly benchmarks (which spawn many java processes with many threads across 128 CPU cores).  It is clearly a deadlock – when it strikes, the process hangs forever until I notice and kill -9 it.  I posted a coupled jstacks in the issue above.

Robert Muir suggested using classcycle to maybe statically dig into possible deadlocks ... I have not tried that yet.

Issue Links

• relates to: LUCENE-9672 Detects deadlock occurring at static initialization timing
• links to: GitHub Pull Request #2200