Windows Compilation VMs are Timing Out After 600 Seconds

Products

• VMware Tanzu Application Service

Issue/Introduction

Unable to apply changes to the Elastic Application Runtime Microsoft Windows add-on due to the compilation VMs timing out. This may also happen when recreating Windows Diego Cells.

Error example:

Task 167201 | 12:30:43 | Compiling packages: golang-1.25-windows/117efdd98cac4d6d36EXAMPLEc5477b5ada0 (00:10:34)
                       L Error: Timed out pinging VM 'compilation-c5e9e60EXAMPLE7-4994_pas-windows-add4EXAMPLEfa33f' with agent '287cbEXAMPLE75852bf2fc' after 600 seconds
Task 167201 | 12:41:17 | Compiling packages: defender_windows/ba8edcEXAMPLEaff1a39cd5927 (00:10:34)
                       L Error: Timed out pinging VM 'compilation-406ec3b1-EXAMPLEe204c_915800410fa5' with agent '8efcd916-EXAMPLE22de8eac93' after 600 seconds
Task 167201 | 12:41:17 | Compiling packages: golang-1-windows/82dfbd5ebcEXAMPLEd6fe2f83ce7f9b1d4 (00:10:34)

Environment

This issue affects Windows stemcells running on vSphere with an NVMe controller

Cause

This issue is caused by NVMe drives not showing unique UUID Serial Number.

Follow these steps to confirm the issue:

1. Power off failed VM in vSphere
2. Attach the boot disk to another VM
3. Mount it to a directory and navigate to "var/vcap/bosh/log/service_wrapper.err.log"
4. Check for errors similar to the following:

   • [main] 2026/07/21 17:42:34 ERROR - Agent exited with error: Running bootstrap: Getting ephemeral disk path: Translating disk ID to disk number failed: Get-Disk : No MSFT_Disk objects found with property 'UniqueId' equal to '6000c29EXAMPLE21be13f4'.  Verify 
     the value of the property and retry.

Resolution

If you are using Windows 2022.x stemcells, you can upgrade the hardware version to 21 or higher and rebuild the stemcell to resolve the issue

If you are using Windows 2019.x stemcells, you will need to rebuild the base VM with a different storage controller (such as SCSI, SATA, etc) and rebuild the stemcell