GitHub actions changes to push windows 2022 docker image to azurecr

[sxa]

Checklist

• commit message has one of the standard prefixes
• faq.md updated if appropriate
• other documentation is changed or added (if applicable)
• playbook changes run through VPC or QPC (if you have access)
• VPC/QPC not applicable for this PR
• for inventory.yml changes, bastillion/nagios/jenkins updated accordingly

[sxa]

Last push is because it looks like win_shell: curl ... invokes the powershell "curl" which doesn't accept the same options:

TASK [ANT : Download ant-contrib with curl] ************************************
task path: /cygdrive/c/infrastructure/ansible/playbooks/AdoptOpenJDK_Windows_Playbook/roles/ANT/tasks/main.yml:66
fatal: [localhost]: FAILED! => {"changed": true, "cmd": "curl -Lo c:\\temp\\ant-contrib.zip https://sourceforge.net/projects/ant-contrib/files/ant-contrib/ant-contrib-1.0b2/ant-contrib-1.0b2-bin.zip", "delta": "0:00:01.704083", "end": "2025-10-13 09:49:16.969501", "msg": "non-zero return code", "rc": 1, "start": "2025-10-13 09:49:15.265418", "stderr": "Invoke-WebRequest : A parameter cannot be found that matches parameter name 'Lo'.\r\nAt line:1 char:70\r\n+ ... InputEncoding = New-Object Text.UTF8Encoding $false; curl -Lo c:\\temp ...\r\n+                                                               ~~~\r\n    + CategoryInfo          : InvalidArgument: (:) [Invoke-WebRequest], ParameterBindingException\r\n    + FullyQualifiedErrorId : NamedParameterNotFound,Microsoft.PowerShell.Commands.InvokeWebReque \r\n   stCommand", "stderr_lines": ["Invoke-WebRequest : A parameter cannot be found that matches parameter name 'Lo'.", "At line:1 char:70", "+ ... InputEncoding = New-Object Text.UTF8Encoding $false; curl -Lo c:\\temp ...", "+                                                               ~~~", "    + CategoryInfo          : InvalidArgument: (:) [Invoke-WebRequest], ParameterBindingException", "    + FullyQualifiedErrorId : NamedParameterNotFound,Microsoft.PowerShell.Commands.InvokeWebReque ", "   stCommand"], "stdout": "", "stdout_lines": []}

[sxa]

Hmmm last GHA run ran out of space:

re-exec error: exit status 1: output: write \\?\C:\Windows\SystemTemp\hcs846643319\Files\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\Llvm\x64\bin\clang-move.exe: There is not enough space on the disk.
Error: Process completed with exit code 1.

[sxa]

(Force push to squash commits and trigger another run to see if it gets the disk space error again)

[sxa]

@gdams Do you think we'd be able to use larger Windows runners to resolve the space issues? I don't believe you had these problems on the earlier PR that this is based on, but perhaps the sizes have changed since then or we're consuming more space now ... I'm not quite sure how that would be charged, and it would need some EF intervention I guess to set them up on the adoptium org.

[sxa]

Noting that with some extra debug in the runners appear to be showing 125GiB free at the start of the playbook runs:

TASK [logs : Run fsutil] *******************************************************
task path: /cygdrive/c/infrastructure/ansible/playbooks/AdoptOpenJDK_Windows_Playbook/roles/logs/tasks/main.yml:46
changed: [localhost] => {"changed": true, "cmd": "fsutil volume diskfree c:", "delta": "0:00:01.437699", "end": "2025-10-14 12:05:53.174145", "rc": 0, "start": "2025-10-14 12:05:51.736445", "stderr": "", "stderr_lines": [], "stdout": "Total free bytes                : 136,073,142,272 (126.7 GB)\r\nTotal bytes                     : 136,229,920,768 (126.9 GB)\r\nTotal quota free bytes          : 136,073,142,272 (126.7 GB)\r\nUnavailable pool bytes          :               0 (  0.0 KB)\r\nQuota unavailable pool bytes    :               0 (  0.0 KB)\r\nUsed bytes                      :     156,581,888 (149.3 MB)\r\nTotal Reserved bytes            :         196,608 (192.0 KB)\r\nVolume storage reserved bytes   :               0 (  0.0 KB)\r\nAvailable committed bytes       :               0 (  0.0 KB)\r\nPool available bytes            :               0 (  0.0 KB)\r\n", "stdout_lines": ["Total free bytes                : 136,073,142,272 (126.7 GB)", "Total bytes                     : 136,229,920,768 (126.9 GB)", "
TASK [logs : Run wmic] *********************************************************
task path: /cygdrive/c/infrastructure/ansible/playbooks/AdoptOpenJDK_Windows_Playbook/roles/logs/tasks/main.yml:49
changed: [localhost] => {"changed": true, "cmd": "wmic logicaldisk get size,freespace,caption", "delta": "0:00:02.314252", "end": "2025-10-14 12:05:58.804099", "rc": 0, "start": "2025-10-14 12:05:56.489847", "stderr": "", "stderr_lines": [], "stdout": "Caption  FreeSpace     Size          \r\r\nC:       136073134080  136229920768  \r\r\n\r\r\n", "stdout_lines": ["Caption  FreeSpace     Size          ", "", "C:       136073134080  136229920768  ", "", "", ""]}
TASK [Debug : Display Debug Information] ***************************************

(I'll update this comment when it runs the same thing at the end of the playbook)

[steelhead31]

Github runners ( for public repos ) appear to only ship with 14Gb of disk space!

https://docs.github.com/en/actions/reference/runners/github-hosted-runners#standard-github-hosted-runners-for-public-repositories

[sxa]

Github runners ( for public repos ) appear to only ship with 14Gb of disk space!

https://docs.github.com/en/actions/reference/runners/github-hosted-runners#standard-github-hosted-runners-for-public-repositories

I saw that too, although the evidence on the machine is that the WINDOWS ones have 125 (I missed the 2 when typing in my earlier comment and wrote it as 15GiB which was incorrect - now edited)

This action run suggests it had 93,398,855,680 ( 87.0 GB) free at the end of the playbook run:

sxa:build-farm$ cat cabbage | sed 's/\\n/\n/g'
task path: /cygdrive/c/infrastructure/ansible/playbooks/AdoptOpenJDK_Windows_Playbook/roles/logs/tasks/main.yml:46
changed: [localhost] => {"changed": true, "cmd": "fsutil volume diskfree c:", "delta": "0:00:01.376376", "end": "2025-10-14 12:34:15.076400", "rc": 0, "start": "2025-10-14 12:34:13.700023", "stderr": "", "stderr_lines": [], "stdout": "Total free bytes                :  93,398,855,680 ( 87.0 GB)\r
Total bytes                     : 136,229,920,768 (126.9 GB)\r
Total quota free bytes          :  93,398,855,680 ( 87.0 GB)\r
Unavailable pool bytes          :               0 (  0.0 KB)\r
Quota unavailable pool bytes    :               0 (  0.0 KB)\r
Used bytes                      :  42,815,275,008 ( 39.9 GB)\r
Total Reserved bytes            :      15,790,080 ( 15.1 MB)\r
Volume storage reserved bytes   :               0 (  0.0 KB)\r
Available committed bytes       :               0 (  0.0 KB)\r
Pool available bytes            :               0 (  0.0 KB)\r

before it goes into a black hole of doing docker stuff for 12 minutes and then falls over. but even if it's copying the entire image around I would expect that to exhaust the remaining 93GiB (Noting that those figures are coming from running the commands inside the container)

[sxa]

Dropping this back to draft just now. I've manually pushed up the image created by https://ci.adoptium.net/job/win2022_docker_image_updater using the azure CLI on that machine:

az account show --query tenantId -o tsv
az login --tenant <id from last command>
az acr login --name adoptium
docker tag win2022_notrhel_image adoptium.azurecr.io/windows2022_build_image
docker push adoptium.azurecr.io/windows2022_build_image