ubuntu-latest: builds hanging in 20260302.42.1 runner image

[iamtesch]

Description

Description

We have a C++ project which the GitHub Action runner compiles (within a docker container, which has been unchanged for months). Normally, this builds fine (and continues to the next step in our workflow). After the latest ubuntu-24.04 image update, the CI build step running the compilation is hanging.

The failed build eventually times out (after no terminal updates about 1 minute into compilation), reporting

The hosted runner lost communication with the server. Anything in your workflow that terminates the runner process, starves it for CPU/Memory, or blocks its network access can cause this error.

We retried the builds with similar effect.

We have verified that the code is unchanged from the working build on the previous runner image.

In previously filed issue #13182, the same symptoms occurred on the same repositories as in the current issue. In this case, several other people had issues with the image and the GitHub actions team resolved the issues by rolling back the changes from that image; I'm not sure if there's anything helpful in there.

Platforms affected

• Azure DevOps
• GitHub Actions - Standard Runners
• GitHub Actions - Larger Runners

Runner images affected

• Ubuntu 22.04
• Ubuntu 24.04
• Ubuntu Slim
• macOS 14
• macOS 14 Arm64
• macOS 15
• macOS 15 Arm64
• macOS 26
• macOS 26 Arm64
• Windows Server 2022
• Windows Server 2025
• Windows Server 2025 with Visual Studio 2026

Image version and build link

20260302.42.1

https://github.com/actions/runner-images/releases/tag/ubuntu24%2F20260302.42

Is it regression?

Yes, this worked a month ago (February 4) on image 20260126.10.1 https://github.com/actions/runner-images/releases/tag/ubuntu24%2F20260126.10

Expected behavior

We expect the build to finish and continue to the next step in our workflow.

Actual behavior

The CI build step running the compilation is hanging when using the most recent ubuntu-latest image.

The failed build eventually times out, reporting

The hosted runner lost communication with the server. Anything in your workflow that terminates the runner process, starves it for CPU/Memory, or blocks its network access can cause this error.

Repro steps

Unfortunately, this is not on a public code repro, but we can reproduce by running the workflow which compiles the code, and I'm happy to follow up and provide more diagnostics or run tests here.

[v-SergeiRadich]

Hi!

We will investigate the issue and reply with results.

[danmoore2205]

We're also experiencing builds hanging for builds that had no issues for months. All they are doing is exporting contentful CMS space data to json for backup purposes. We run it nightly on 3 different spaces and have been seeing intermittent failures and our largest space has not worked since March 2nd.

https://www.contentful.com/developers/docs/tutorials/cli/import-and-export/ I tried using 22.04 instead of 24.04 but the issue persisted

[t0m-4]

Hello,

We have been experiencing exactly the same error in our pipeline for a few days now, even though it had been running without any issues for months.

Our Azure DevOps pipeline contains 3 stages, and each stage has jobs that themselves contain tasks.

Here is roughly the configuration:

- stage: Configuration
  - job: ComputeVersionJob
    ...

  - job: ConfigureBuildActionsJob
    steps:
      # No checkout needed
      - checkout: self
        clean: true
        fetchFilter: tree:0 # no file contents, just commit history
        sparseCheckoutPatterns: # fake pattern to force git to download nothing just persist credentials
          - /__nothing_matches__/
        persistCredentials: true

      - task: PowerShell@2
         ...

      - task: PowerShell@2
        # use git diff-tree command
        script: |
          git diff-tree --no-commit-id --name-only -r "$prLastMergeTargetCommit..$commitId"

The last PowerShell task runs for about 1 hour with the following message:

##[warning]Free memory is lower than 5%; Currently used: 95.55%

Finally, the job ends with the following error:

##[error]We stopped hearing from agent Azure Pipelines. Verify the agent machine is running and has a healthy network connection. Anything that terminates an agent process, starves it for CPU, or blocks its network access can cause this error. For more information, see: https://go.microsoft.com/fwlink/?linkid=846610

There are only two files modified between the last two commits.
Thanks in advance.

[iamtesch]

Additional information on our run -- we are using a custom docker image in which we build. If this is a memory starvation issue, is there a way to understand how much memory is available on each runner image to see if this has changed with a runner image update, and is there a minimum amount that we can rely on moving forward?

[iamtesch]

More confirmation that it seems to be a reduction in memory on the latest image -- we ran with a customized 8-core GitHub runner on ubuntu-latest, and the build worked fine there.

[t0m-4]

Hello,

We have been experiencing exactly the same error in our pipeline for a few days now, even though it had been running without any issues for months.

Our Azure DevOps pipeline contains 3 stages, and each stage has jobs that themselves contain tasks.

Here is roughly the configuration:

• stage: Configuration

  • job: ComputeVersionJob
    ...

  • job: ConfigureBuildActionsJob
    steps:

    No checkout needed

    • checkout: self
      clean: true
      fetchFilter: tree:0 # no file contents, just commit history
      sparseCheckoutPatterns: # fake pattern to force git to download nothing just persist credentials

      • /nothing_matches/
        persistCredentials: true

    • task: PowerShell@2
      ...

    • task: PowerShell@2

      use git diff-tree command

      script: |
      git diff-tree --no-commit-id --name-only -r "$prLastMergeTargetCommit..$commitId"
      The last PowerShell task runs for about 1 hour with the following message:

##[warning]Free memory is lower than 5%; Currently used: 95.55%

Finally, the job ends with the following error:

##[error]We stopped hearing from agent Azure Pipelines. Verify the agent machine is running and has a healthy network connection. Anything that terminates an agent process, starves it for CPU, or blocks its network access can cause this error. For more information, see: https://go.microsoft.com/fwlink/?linkid=846610

There are only two files modified between the last two commits. Thanks in advance.

Hello got some updates,

I tried with ubuntu 22.04 same issue then i tried with windows 2025 image

Get files updated between commits 9ce130a07aae5b5830411665d1656856ca17c0bf and 3ec4312ccb4e5a45c9f5924664cacab6cc30f116
fatal: full write to remote helper failed: Broken pipe
fatal: could not fetch 5da6d957b5fff9986916d75ade1b4a8b61bc78a1 from promisor remote
fatal: fetch-pack: invalid index-pack output
fatal: could not fetch 5da6d957b5fff9986916d75ade1b4a8b61bc78a1 from promisor remote
fatal: fetch-pack: invalid index-pack output
fatal: could not fetch 5da6d957b5fff9986916d75ade1b4a8b61bc78a1 from promisor remote
...
fatal: Out of memory, malloc failed (tried to allocate 217 bytes)
fatal: the remote end hung up unexpectedly

[v-GeorgyPuzakov]

@iamtesch Are you still experiencing this issue?