github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-24 14:43:39 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #15597] ollama ls takes 5 seconds in interactive terminal (pselect6 timeout on stdout) #87642

New Issue
Open
opened 2026-05-10 06:11:58 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-10 06:11:58 -05:00
Owner
Copy Link

Originally created by @felix021 on GitHub (Apr 15, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/15597

Describe the bug

ollama ls (and other list/ps commands) consistently takes exactly 5 seconds when run in an interactive terminal (PTY). The same command completes in ~14ms when stdout is piped or not connected to a TTY.

Steps to reproduce

  1. Run time ollama ls in any interactive terminal (xterm, tmux, gnome-terminal, etc.)
  2. Observe ~5 second delay before output appears

Reproducible via script which creates a PTY:

# Slow (5s) - with PTY
script -qc 'time ollama ls' /dev/null

# Fast (14ms) - without PTY
time ollama ls | cat

Expected behavior

ollama ls should complete in milliseconds regardless of terminal type.

Root cause (strace analysis)

strace -f -T shows the TUI rendering framework performs a pselect6 on stdout (fd 1) with a 5-second timeout:

pselect6(2, [1], NULL, NULL, {tv_sec=5, tv_nsec=0}, NULL) = 1 (in [1])       # first call succeeds
pselect6(2, [1], NULL, NULL, {tv_sec=5, tv_nsec=0}, NULL) = 0 (Timeout) <5.005162>  # second call times out

After the 5s timeout, the process proceeds normally:

ioctl(1, TCSETS, ...)  # terminal control
SIGURG                  # Go runtime signal

The 5-second wall time is entirely I/O wait (user=7ms, sys=14ms, real=5020ms).

Workaround

Setting TERM=dumb bypasses the TUI rendering and eliminates the delay:

TERM=dumb time ollama ls  # 14ms

Alternatively, pipe output: ollama ls | cat

Environment

  • OS: Ubuntu 24.04, Linux 6.8.0-107-generic
  • Ollama: v0.18.2
  • GPU: NVIDIA GeForce RTX 5060 Ti (driver 590.48.01, CUDA 13.1)
  • Terminal: Reproduced in gnome-terminal, tmux, and via script -qc
  • Shell: bash 5.2

Additional context

The issue is not related to network/proxy configuration (verified by testing with and without proxy). It is purely a terminal I/O issue in the CLI's TUI framework (likely bubbletea). The pselect6 on stdout fd behavior suggests a race condition or incorrect fd readiness check in the terminal rendering pipeline.

Originally created by @felix021 on GitHub (Apr 15, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/15597 ## Describe the bug `ollama ls` (and other list/ps commands) consistently takes exactly **5 seconds** when run in an interactive terminal (PTY). The same command completes in ~14ms when stdout is piped or not connected to a TTY. ## Steps to reproduce 1. Run `time ollama ls` in any interactive terminal (xterm, tmux, gnome-terminal, etc.) 2. Observe ~5 second delay before output appears Reproducible via `script` which creates a PTY: ```bash # Slow (5s) - with PTY script -qc 'time ollama ls' /dev/null # Fast (14ms) - without PTY time ollama ls | cat ``` ## Expected behavior `ollama ls` should complete in milliseconds regardless of terminal type. ## Root cause (strace analysis) `strace -f -T` shows the TUI rendering framework performs a `pselect6` on stdout (fd 1) with a 5-second timeout: ``` pselect6(2, [1], NULL, NULL, {tv_sec=5, tv_nsec=0}, NULL) = 1 (in [1]) # first call succeeds pselect6(2, [1], NULL, NULL, {tv_sec=5, tv_nsec=0}, NULL) = 0 (Timeout) <5.005162> # second call times out ``` After the 5s timeout, the process proceeds normally: ``` ioctl(1, TCSETS, ...) # terminal control SIGURG # Go runtime signal ``` The 5-second wall time is entirely I/O wait (`user=7ms, sys=14ms, real=5020ms`). ## Workaround Setting `TERM=dumb` bypasses the TUI rendering and eliminates the delay: ```bash TERM=dumb time ollama ls # 14ms ``` Alternatively, pipe output: `ollama ls | cat` ## Environment - **OS**: Ubuntu 24.04, Linux 6.8.0-107-generic - **Ollama**: v0.18.2 - **GPU**: NVIDIA GeForce RTX 5060 Ti (driver 590.48.01, CUDA 13.1) - **Terminal**: Reproduced in gnome-terminal, tmux, and via `script -qc` - **Shell**: bash 5.2 ## Additional context The issue is not related to network/proxy configuration (verified by testing with and without proxy). It is purely a terminal I/O issue in the CLI's TUI framework (likely bubbletea). The `pselect6` on stdout fd behavior suggests a race condition or incorrect fd readiness check in the terminal rendering pipeline.
GiteaMirror commented 2026-05-10 06:11:59 -05:00
Author
Owner
Copy Link

@PureBlissAK commented on GitHub (Apr 18, 2026):

🤖 Automated Triage & Analysis Report

Issue: #15597
Analyzed: 2026-04-18T18:19:33.608983

Analysis

  • Type: unknown
  • Severity: medium
  • Components: unknown

Implementation Plan

  • Effort: medium
  • Steps:

This issue has been triaged and marked for implementation.

<!-- gh-comment-id:4274305108 --> @PureBlissAK commented on GitHub (Apr 18, 2026): <!-- ollama-issue-orchestrator:v1 issue:15597 --> ## 🤖 Automated Triage & Analysis Report **Issue**: #15597 **Analyzed**: 2026-04-18T18:19:33.608983 ### Analysis - **Type**: unknown - **Severity**: medium - **Components**: unknown ### Implementation Plan - **Effort**: medium - **Steps**: *This issue has been triaged and marked for implementation.*
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#87642
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?