Skip to content

fix: bump pinned Ollama version to 0.32.2 (rc0 tag removed)#1425

Merged
planetf1 merged 1 commit into
generative-computing:mainfrom
planetf1:fix/ollama-version-pin
Jul 22, 2026
Merged

fix: bump pinned Ollama version to 0.32.2 (rc0 tag removed)#1425
planetf1 merged 1 commit into
generative-computing:mainfrom
planetf1:fix/ollama-version-pin

Conversation

@planetf1

Copy link
Copy Markdown
Contributor

Pull Request

Issue

Fixes #1424

Description

.github/workflows/quality.yml pins OLLAMA_VERSION=0.32.2-rc0 (added in 39002da to fix the #1388 GCC-13 AMX segfault). Ollama has since promoted that release candidate to the v0.32.2 release and removed the rc0 tag, so install.sh 404s on the missing version and the Install Ollama step fails on every pull_request and merge_group run — blocking all PRs and the merge queue.

This bumps the pin to 0.32.2:

  • Verified downloadable: releases/download/v0.32.2/ollama-linux-amd64.tar.zst returns HTTP 200; the v0.32.2-rc0 asset returns 404.
  • Pin kept, not dropped: dropping it would fall back to the latest stable release, 0.32.1, which is the GCC-11 build that reintroduces bug(ci): llama-server segfault cascade in quality job #1388. v0.32.2 carries the same GCC-13 build as the rc.
  • The comment is updated to record the rc0 → release promotion; the guidance to drop the pin once a stable (non-prerelease) release includes the GCC-13 build is retained.

This PR's own quality run is the validation: a green run confirms 0.32.2 installs cleanly.

Testing

  • Tests added to the respective file if code was changed
  • New code has 100% coverage if code was added
  • Ensure existing tests and github automation passes — CI workflow change; validated by this PR's own quality run

Attribution

  • AI coding assistants used

Adding a new component, requirement, sampling strategy, or tool?

If your PR adds or modifies one of the types below, check the matching box. A checklist of type-specific review items will be posted as a comment.

  • Component
  • Requirement
  • Sampling Strategy
  • Tool

NOTE: Please ensure you have an issue that has been acknowledged by a core contributor and routed you to open a pull request against this repository. Otherwise, please open an issue before continuing with this pull request.

The quality workflow pins OLLAMA_VERSION=0.32.2-rc0 (added to fix the generative-computing#1388
GCC-13 AMX segfault). Ollama has since promoted that release candidate to the
v0.32.2 release and removed the rc0 tag, so install.sh now 404s on the missing
version and the "Install Ollama" step fails on every PR and merge-queue run.

Bump the pin to 0.32.2, whose linux-amd64 asset is present and which carries
the same GCC-13 build. The pin is kept (not dropped) because dropping it falls
back to the latest stable release, 0.32.1, which is the GCC-11 build that
reintroduces generative-computing#1388.

Assisted-by: Claude Code
Signed-off-by: Nigel Jones <jonesn@uk.ibm.com>
@planetf1
planetf1 enabled auto-merge July 22, 2026 11:26
@planetf1
planetf1 added this pull request to the merge queue Jul 22, 2026
Merged via the queue into generative-computing:main with commit 6c96da3 Jul 22, 2026
9 checks passed
@planetf1
planetf1 deleted the fix/ollama-version-pin branch July 22, 2026 12:28
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug Something isn't working

Projects

None yet

Development

Successfully merging this pull request may close these issues.

CI: every PR blocked — Install Ollama step fails on removed 0.32.2-rc0 tag

2 participants