Public issue detail

Runaway CPU investigation for llama-server: busy poll at ggml_backend_sched_split_graph

llama-server is stuck in a likely busy poll loop: 3.40% of sampled CPU passed through ggml_backend_sched_split_graph, with repeated thread backtraces show 22 thread(s) around 0x00007f942aaa74be in ?? () from target:<path> and 1 thread(s) around 0x00007f942aaa74be in ?? () from target:<path>.

highscore: 106reports: 1

Last seen: 2026-07-08 23:45 UTC. Public JSON: /v1/issues/019f43fd-8ab8-7533-b8b0-1c9ebacecb50

Technical snapshot

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

Representative thread backtrace

  • Command: /usr/local/lib/ollama/llama-server --model /usr/share/ollama/.ollama/models/blobs/sha256-797b70c4edf85907fe0a49eb85811256f65fa0f7bf52166b147fd16be2be4662 --port 33809 --host 127.0.0.1 --no-webui --offline -c 512 -np 1 --log-verbosity 4 --no-log-prefix --no-log-timestamps --flash-attn auto --embedding -b 512 -ub 512 -t 8 --context-shift --keep 4
  • Why Fixer classified it this way: The trace repeatedly returns to a poll-family syscall without meaningful blocking, which suggests a busy event-loop wakeup.
  • Thread summary: thread backtraces show 22 thread(s) around 0x00007f942aaa74be in ?? () from target:<path> and 1 thread(s) around 0x00007f942aaa74be in ?? () from target:<path>
  • Contention signals: gdb-stderr: gdb: warning: Couldn't determine a path for the index cache directory.
  • Repeated loop: futex -> futex -> futex
  • Top syscalls: futex x1949, <... futex resumed>) = -1 ETIMEDOUT x274, poll x237, setsockopt x183
  • Kernel: 6.17.10+deb14-amd64
  • Distribution: debian
0x00007f942aaa74be in ?? () from target:<path>
0x00007f942aa9bc84 in ?? () from target:<path>
0x00007f942aa9c29c in ?? () from target:<path>
0x00007f942aa9e828 in pthread_cond_wait () from target:<path>
0x00007f942acbaaa7 in std::condition_variable::wait(std::unique_lock<std::mutex>&) () from target:<path>++.so.6
0x00007f942b65c90b in std::thread::_State_impl<std::thread::_Invoker<std::tuple<common_log::resume()::{lambda()#1}> > >::_M_run() () from target:<path>
0x00007f942acec446 in ?? () from target:<path>++.so.6
0x00007f942aa9f289 in ?? () from target:<path>

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.

Runaway CPU investigation for ollama: busy poll at fdget

possible duplicate

ollama is stuck in a likely busy poll loop: 27.49% of sampled CPU passed through fdget, with repeated thread backtraces show 2 thread(s) around 0x00007f9c0b4a74be in ?? () from target:<path> and 1 thread(s) around 0x000055cccbf41143 in ?? ().

highscore: 106reports: 1successful triagesimilarity: 79%

Why this looks related: same classification, same subsystem, same wait site

Last seen: 2026-07-10 03:00 UTC. Public page: /issues/019dbea3-2751-7eb3-9814-524b0b2e18f6. Public JSON: /v1/issues/019dbea3-2751-7eb3-9814-524b0b2e18f6

Runaway CPU investigation for ollama: busy poll at check_stack_object

possible duplicate

ollama is stuck in a likely busy poll loop: 25.00% of sampled CPU passed through check_stack_object, with repeated thread backtraces show 2 thread(s) around 0x00007f5ea9bb14be in ?? () from target:<path> and 1 thread(s) around 0x000055b0b942d143 in ?? ().

highscore: 106reports: 1similarity: 77%

Why this looks related: same classification, same subsystem, same wait site

Last seen: 2026-07-08 11:50 UTC. Public page: /issues/019f415e-58c4-7333-8af7-39977d9e0c41. Public JSON: /v1/issues/019f415e-58c4-7333-8af7-39977d9e0c41

Runaway CPU investigation for ollama: busy poll at __get_pfnblock_flags_mask.isra.0

possible duplicate

ollama is stuck in a likely busy poll loop: 33.33% of sampled CPU passed through __get_pfnblock_flags_mask.isra.0, with repeated thread backtraces show 2 thread(s) around 0x00007f9274ca74be in ?? () from target:<path> and 1 thread(s) around 0x000055d3133cc143 in ?? ().

highscore: 106reports: 1successful triagesimilarity: 73%

Why this looks related: same classification, same subsystem, same wait site

Last seen: 2026-07-08 10:37 UTC. Public page: /issues/019e5346-afaf-7142-8a8d-fc9713f903ce. Public JSON: /v1/issues/019e5346-afaf-7142-8a8d-fc9713f903ce

Runaway CPU investigation for llama-server: timer churn at mt_find

possible duplicate

llama-server is stuck in a likely timer churn loop: 8.42% of sampled CPU passed through mt_find, with repeated thread backtraces show 22 thread(s) around 0x00007f03399b14be in ?? () from target:<path> and 1 thread(s) around 0x00007f03399b14be in ?? () from target:<path>.

highscore: 106reports: 1similarity: 73%

Why this looks related: same subsystem, same target, same wait site

Last seen: 2026-07-09 01:53 UTC. Public page: /issues/019f4473-347f-75d0-8ee4-58cc8558c55b. Public JSON: /v1/issues/019f4473-347f-75d0-8ee4-58cc8558c55b

Runaway CPU investigation for llama-server: unknown userspace loop at num_to_str

possible duplicate

llama-server is stuck in a likely unclassified userspace loop: 10.95% of sampled CPU passed through num_to_str, with repeated thread backtraces show 22 thread(s) around 0x00007f942aaa74be in ?? () from target:<path> and 1 thread(s) around 0x00007f942aaa74be in ?? () from target:<path>.

highscore: 106reports: 1similarity: 71%

Why this looks related: same subsystem, same target, same wait site

Last seen: 2026-07-08 20:41 UTC. Public page: /issues/019f4355-227b-7972-b32b-3ce63462b54c. Public JSON: /v1/issues/019f4355-227b-7972-b32b-3ce63462b54c

Runaway CPU investigation for llama-server: unknown userspace loop at mt_find

possible duplicate

llama-server is stuck in a likely unclassified userspace loop: 11.72% of sampled CPU passed through mt_find, with repeated thread backtraces show 22 thread(s) around 0x00007f1cf9ca74be in sem_trywait () from target:<path> and 1 thread(s) around 0x00007f1cf9ca74be in sem_trywait () from target:<path>.

highscore: 106reports: 1similarity: 69%

Why this looks related: same subsystem, same target, same wait site

Last seen: 2026-07-10 03:00 UTC. Public page: /issues/019f427c-2154-77c0-818c-eded7911a06a. Public JSON: /v1/issues/019f427c-2154-77c0-818c-eded7911a06a

Published attempts

No public attempts have been published for this issue yet.

Quarantine stays on by default so the public queue is harder to spam.