Description
Description
In the past two weeks, one of our build pipelines is failing on the steps to run tests because it cannot access SQL Server container created in the previous step. The report back from the container is the same as it was before the job failed, and since we use our own ACR, the image had not been updated in-between the job previously succeeding and now failing: Container Details: [ContainerID] mssql-windows-mssql-server-windows Up 7 seconds (health: starting) 0.0.0.0:1433->1433/tcp
The next step in this job is to try and connect to the port to ensure the container is running and accessible, however this now fails. Because of the way the pipeline works the next job begins which would run the tests, and that is also unable to connect to the container.
Platforms affected
- Azure DevOps
- GitHub Actions - Standard Runners
- GitHub Actions - Larger Runners
Runner images affected
- Ubuntu 22.04
- Ubuntu 24.04
- macOS 13
- macOS 13 Arm64
- macOS 14
- macOS 14 Arm64
- macOS 15
- macOS 15 Arm64
- Windows Server 2019
- Windows Server 2022
- Windows Server 2025
Image version and build link
20250511.1.0
Is it regression?
Yes, 20250427.1.0
Expected behavior
The TCP/IP connection to the container succeeds and tests are able to run.
Actual behavior
The container is unresponsive or the port is inaccessible, and the connection test and the tests themselves are unable to connect.
Repro steps
- Create a container in an ADO pipeline using Powershell.
- Try to connect using TCP/IP to the exposed port.