Error: getCacheEntry failed: Cache service responded with 500 in new workflow

Description: setup-node fails with an error

Run actions/setup-node@v2
  with:
    node-version: 16
    cache: pnpm
    cache-dependency-path: **/pnpm-lock.yaml
    always-auth: false
    check-latest: false
    token: ***
Found in cache @ /opt/hostedtoolcache/node/16.13.0/x64
/opt/hostedtoolcache/node/16.13.0/x64/bin/pnpm store path
/home/runner/.pnpm-store/v3
Error: getCacheEntry failed: Cache service responded with 500

Action version: setup-node@v2

Platform:

• Ubuntu
• macOS
• Windows

Runner type:

• Hosted
• Self-hosted

Tools version: node 14 + 16 pnpm 6.23

Repro steps:
Workflow here https://github.com/svitejs/vite-plugin-env-config/blob/main/.github/workflows/ci.yml

It uses setup-node twice:

• first to get npm and run npm i -g pnpm (instead of using pnpms custom action) https://github.com/svitejs/vite-plugin-env-config/blob/9fc3bbddc0b93583a3f068e5f0b8a6ecc0275a0a/.github/workflows/ci.yml#L27

• second with pnpm cache option set to use automatic caching https://github.com/svitejs/vite-plugin-env-config/blob/9fc3bbddc0b93583a3f068e5f0b8a6ecc0275a0a/.github/workflows/ci.yml#L31

Expected behavior:

setup succeeds

Actual behavior: setup fails log shows a 500 error for cache service https://github.com/svitejs/vite-plugin-env-config/runs/4494014103?check_suite_focus=true#step:5:12

Additional information

• The repo is very new and cloned from a template. other repos with a similar setup don’t show this behavior
• These are the initial workflow runs, there hasn’t been a success. Maybe some initializaion failed?

[gregnrobinson]

I am also seeing similar behaviour in my workflows…

https://github.community/t/getcacheentry-failed-id-1-innerexception-null-message-database-pool-defaultpartitionpool-is-full-raise-the-maxdatabaselimit-or-use-a-different-pool/217152

https://github.com/gregnrobinson/kube-secrets/runs/4498019768?check_suite_focus=true

[dmitry-shibanov]

Hello everyone. Thank you for your report. I’m going to close the issue, because the initial one was resolved. If you have any concerns feel free to ping us.