github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #1569] [CLOSED] RFC - initial foundation for distributed integration test rigging #10619

New Issue
Closed
opened 2026-04-12 23:05:13 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-12 23:05:13 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/1569
Author: @dhiltgen
Created: 12/16/2023
Status: ❌ Closed

Base: main ← Head: integration_testing

📝 Commits (1)

  • 15c2d4b First cut at some distributed test rigging

📊 Changes

13 files changed (+526 additions, -16 deletions)

View changed files

📝 .dockerignore (+1 -1)
📝 .gitignore (+2 -1)
📝 Dockerfile.build (+2 -0)
➕ scripts/binaries.py (+67 -0)
➕ scripts/build_all.sh (+63 -0)
📝 scripts/build_darwin.sh (+23 -8)
📝 scripts/build_linux.sh (+1 -0)
📝 scripts/build_remote.py (+27 -1)
➕ scripts/machine.py (+224 -0)
➕ scripts/regression_suite.py (+62 -0)
📝 server/llm_image_test.go (+2 -2)
📝 server/llm_test.go (+6 -3)
➕ test-systems.yaml (+46 -0)

📄 Description

This adds some new script rigging that may help flesh out integration tests.

It doesn't actually do any interesting testing yet (just a hello world), but demonstrates how this could work and be layered into PyTest or similar frameworks.

At present, it requires you're running on a mac.
First build all target architectures:

./scripts/build_all.sh <some-windows-remote-name>

This will produce both regular and coverage instrumented binaries in ./dist/

There's a yaml config file to drive what the remote target systems are: test-system.yaml

Then you can run a

python3 ./scripts/regression_suite.py

which will figure out the remote target system OS/arch, and then copy over the required binary (if not already present) and then run the server with coverage, and then run a hello world client, then shutdown the server, and gather coverage data. You can use the experimental variant env var to force certain llm libraries to increase coverage.

I've confirmed it works on linux and windows.

Notable gaps:

  • Still need to add a docker/container based test mechanism
  • Serial running is lame, but I just wanted to prove out the concept - it should really use some test framework to parallelize the run across all the systems, and be smart about hostnames so it doesn't hit the same host multiple times in parallel.
  • WSL on windows needs support
  • It should work on remote mac systems, but I haven't tested it yet.

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/1569 **Author:** [@dhiltgen](https://github.com/dhiltgen) **Created:** 12/16/2023 **Status:** ❌ Closed **Base:** `main` ← **Head:** `integration_testing` --- ### 📝 Commits (1) - [`15c2d4b`](https://github.com/ollama/ollama/commit/15c2d4be7aed27abb19d82f258adfcc58bf01b3c) First cut at some distributed test rigging ### 📊 Changes **13 files changed** (+526 additions, -16 deletions) <details> <summary>View changed files</summary> 📝 `.dockerignore` (+1 -1) 📝 `.gitignore` (+2 -1) 📝 `Dockerfile.build` (+2 -0) ➕ `scripts/binaries.py` (+67 -0) ➕ `scripts/build_all.sh` (+63 -0) 📝 `scripts/build_darwin.sh` (+23 -8) 📝 `scripts/build_linux.sh` (+1 -0) 📝 `scripts/build_remote.py` (+27 -1) ➕ `scripts/machine.py` (+224 -0) ➕ `scripts/regression_suite.py` (+62 -0) 📝 `server/llm_image_test.go` (+2 -2) 📝 `server/llm_test.go` (+6 -3) ➕ `test-systems.yaml` (+46 -0) </details> ### 📄 Description This adds some new script rigging that may help flesh out integration tests. It doesn't actually do any interesting testing yet (just a [hello world](https://github.com/jmorganca/ollama/pull/1569/files#diff-aea44eb9fc093044f1591164d21a64fef7e13e504088f1599e4685f9835ddb4d)), but demonstrates how this could work and be layered into PyTest or similar frameworks. At present, it requires you're running on a mac. First [build all target architectures](https://github.com/jmorganca/ollama/pull/1569/files#diff-ad855df3d2b5ffdf734facd7dbd1c5d4dd1da6fcd50110f86015e64ef9c9d88e): ``` ./scripts/build_all.sh <some-windows-remote-name> ``` This will produce both regular and coverage instrumented binaries in `./dist/` There's a yaml config file to drive what the remote target systems are: [test-system.yaml](https://github.com/jmorganca/ollama/pull/1569/files#diff-e4638e57348bd44ab998a6ec8e425553e91e72712b8da5dafaa2506ac8633408) Then you can run a ``` python3 ./scripts/regression_suite.py ``` which will figure out the remote target system OS/arch, and then copy over the required binary (if not already present) and then run the server with coverage, and then run a hello world client, then shutdown the server, and gather coverage data. You can use the experimental variant env var to force certain llm libraries to increase coverage. I've confirmed it works on linux and windows. Notable gaps: * Still need to add a docker/container based test mechanism * Serial running is lame, but I just wanted to prove out the concept - it should really use some test framework to parallelize the run across all the systems, and be smart about hostnames so it doesn't hit the same host multiple times in parallel. * WSL on windows needs support * It should work on remote mac systems, but I haven't tested it yet. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-12 23:05:13 -05:00
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 pull-request
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#10619
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?