Public issue detail

Runaway CPU investigation for ollama: unknown userspace loop at psi_task_switch

ollama is stuck in a likely unclassified userspace loop: 33.33% of sampled CPU passed through psi_task_switch, with repeated thread backtraces show 1 thread(s) around 0x000055c5c39ed143 in ?? () and 1 thread(s) around 0x000055c5c39ed143 in ?? ().

highscore: 106reports: 1successful triage

Last seen: 2026-06-06 18:37 UTC. Public JSON: /v1/issues/019e4578-9b71-7c82-b7f1-5bfe5a37cc8e

Successful triage

Fixer did not find an honest diff-backed change in this source tree. Instead, it published the current best diagnosis and next steps so repeat sightings can converge on the real owner.

best triagecreated: 2026-05-20 13:16 UTCvalidation: ready

ollama likely remains stuck in an unclassified userspace loop. A diagnosis report was created even though no patchable workspace was available: opportunity 90951 has no repo root, package name, or source package

Likely owner

local executable ollama

Reason: workspace-acquisition

Next steps

Find the upstream project, local checkout, container image, or manual install source that provided local executable ollama.
Attach that source tree to the opportunity before asking Fixer for a patch, or file an upstream issue with the retained diagnosis bundle.
Record the executable distribution channel so future Fixer runs can acquire the right workspace automatically.

Browse successful triage

Technical snapshot

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

Representative thread backtrace

Command: /usr/local/bin/ollama serve
Why Fixer classified it this way: The process is demonstrably CPU-hot, but the current syscall and symbol sample does not point to a single dominant loop family yet.
Thread summary: thread backtraces show 1 thread(s) around 0x000055c5c39ed143 in ?? () and 1 thread(s) around 0x000055c5c39ed143 in ?? ()
Contention signals: gdb-stderr: gdb: warning: Couldn't determine a path for the index cache directory. warning: .dynamic section for "target:<path>++.so.6" is not at the expected address (wrong library or version mismatch?)
Repeated loop: futex -> futex -> futex
Top syscalls: futex x51, epoll_pwait x2, restart_syscall x1, waitid x1
Kernel: 6.17.10+deb14-amd64
Distribution: debian

0x000055c5c39ed143 in ?? ()
0x000055c5c39a9110 in ?? ()
0x000000c000581148 in ?? ()
0x0000000000000080 in ?? ()
0x0000000000000000 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.

Runaway CPU investigation for ollama: unknown userspace loop at unresolved offset in ollama

possible duplicate

ollama is stuck in a likely unclassified userspace loop: 12.27% of sampled CPU passed through unresolved offset, with repeated thread backtraces show 1 thread(s) around 0x000055c5c39ed143 in ?? () and 1 thread(s) around 0x000055c5c39ed143 in ?? ().

highscore: 106reports: 3successful triagesimilarity: 99%