Public issue detail

Runaway CPU investigation for python3.13: futex lock contention at _PyEval_EvalFrameDefault

python3.13 is stuck in a likely futex lock contention loop: 5.90% of sampled CPU passed through _PyEval_EvalFrameDefault, with repeated thread backtraces show 2 thread(s) around 0x00007fd097c7c9ee in futex_fatal_error () from target:<path> and 1 thread(s) around 0x00007fd097c7c9ee in futex_fatal_error () from target:<path>.

highpackage: python3.13-minimalsource: python3.13score: 106reports: 1

Last seen: 2026-11-16 20:23 UTC. Public JSON: /v1/issues/01a18e1b-e88d-7622-b075-af7da145ec9a

Technical snapshot

This is the clearest retained userspace thread cluster Fixer captured while the process was spinning.

Representative thread backtrace

  • Command: python3 ../../utils/docs/postgis_exampletest.py --run --database postgis_exampletest ../doc/postgis-out.xml --keep-going --visual-only --jobs 2 --render-dir /woodpecker/src/gitea.osgeo.org/postgis/postgis-security/build-docs/doc/html/images/visual-examples
  • Why Fixer classified it this way: Multiple captured threads converge on futex or pthread mutex paths, which suggests user-space lock contention or a lock convoy.
  • Thread summary: thread backtraces show 2 thread(s) around 0x00007fd097c7c9ee in futex_fatal_error () from target:<path> and 1 thread(s) around 0x00007fd097c7c9ee in futex_fatal_error () from target:<path>
  • Contention signals: futex-lock-contention, event-loop-wakeups, gdb-stderr: gdb: warning: Couldn't determine a path for the index cache directory. warning: Target and debugger are in different PID namespaces; thread lists and other data are likely unreliable. Connect to gdbserver inside the container. warning: Expected absolute pathname for libpthread in the inferior, but got target:<path> warning: Unable to find libthread_db matching inferior's thread library, thread debugging will not be available.
  • Repeated loop: rt_sigaction -> rt_sigaction -> rt_sigaction
  • Top syscalls: mmap x1800, rt_sigaction x1338, rt_sigprocmask x1125, close x703
  • Package: python3.13-minimal 3.13.15-1
  • Kernel: 7.1.8+deb14.1-amd64
  • Distribution: debian
0x00007fd097c7c9ee in futex_fatal_error () from target:<path>
0x00007fd097c71668 in __GI__IO_flush_all () from target:<path>
0x00007fd097c716ad in __GI__IO_flush_all () from target:<path>
0x00007fd097ce59c6 in poll () from target:<path>
0x0000000000665f4f in ?? ()
0x0000000000562471 in _PyEval_EvalFrameDefault ()
0x00000000005cd343 in ?? ()
0x00000000006dc6a0 in ?? ()

Possible duplicates

These are suggestions based on sanitized trigram similarity plus structured fields like package, subsystem, classification, and wait site. They are not auto-merged.

python3.13 is stuck in a likely busy poll loop: 50.00% of sampled CPU passed through _PyEval_FrameClearAndPop, with repeated thread backtraces show 1 thread(s) around 0x00007f1c6e7efe92 in pthread_attr_destroy () from target:<path>.

highpackage: python3.13-minimalsource: python3.13score: 106reports: 2successful triagesimilarity: 80%

Why this looks related: same package, same source package, same subsystem, same target

Last seen: 2026-11-05 17:49 UTC. Public page: /issues/019dd062-04a8-7e03-b3a3-5bb273aa2f42. Public JSON: /v1/issues/019dd062-04a8-7e03-b3a3-5bb273aa2f42

Published attempts

No public attempts have been published for this issue yet.