Description of problem: AWS EC2 metadata service not available in host's vNIC for Windows after hybrid-overlay is configured and running as a Windows service. Version-Release number of selected component (if applicable): WMCO 5.0.1 How reproducible: Always Steps to Reproduce: 1. Spin a 4.10 IPI cluster in AWS 2. Install WMCO from Operator-Hub with all the requirements 3. Deploy a Windows machineSet with Windows Server 2022 4. SSH into a Windows machine and check for the metadata endpoint is reachable before the machine joins as a worker node. 4.1 Run the following scripts from an existing linux worker node ``` # query the metadata service to fetch the hostname ssh -o StrictHostKeyChecking=no -i <key.pem> <username>@<internal_IP> powershell.exe -Command "Invoke-WebRequest -UseBasicParsing http://169.254.169.254/latest/meta-data/hostname" ``` ``` # get network routes and look for 169.254.169.254 ssh -o StrictHostKeyChecking=no -i <key.pem> <username>@<internal_IP> powershell.exe -Command "Get-NetRoute" ``` where <key.pem> is the private key to SSH into the Windows machine, <username> is the Windows username, and <internal_IP is the internal IP of the Windows machine. 5. Wait for WMCO to start hybrid-overlay as a Windows service 6. SSH into a Windows machine and check again the metadata endpoint Actual results: AWS EC2 metadata service is not available in host's vNIC after hybrid-overlay is running as a Windows service. Expected results: AWS EC2 metadata service should be available in host's vNIC after hybrid-overlay running as a Windows service. Additional info:
Upstream PR with the proposed fix: https://github.com/ovn-org/ovn-kubernetes/pull/3074 Downstream merge PR: https://github.com/openshift/ovn-kubernetes/pull/1214
Moving to POST state; assigning to jvaldes who is the author of PR.
Downstream merge PR merged. https://github.com/openshift/ovn-kubernetes/pull/1214
As a workaround, you can SSH into the Windows instance first, switch to PowerShell and then invoke the command. As an example and based on the above output, the steps are: 1- SSH into the Windows instance ssh -o StrictHostKeyChecking=no -i ./openshift-qe.pem Administrator.compute.internal 2- Change to PowerShell C:\Users\Administrator> powershell.exe 3- Invoke the command directly in PowerShell PS C:\Users\Administrator> Invoke-WebRequest -UseBasicParsing http://169.254.169.254/latest/meta-data/hostname
Verified on 4.12.0-0.nightly-2022-09-20-040107 using steps in comment 5 PS C:\Users\Administrator> Invoke-WebRequest -UseBasicParsing http://169.254.169.254/latest/meta-data/hostname StatusCode : 200 StatusDescription : OK Content : ip-10-0-132-226.us-east-2.compute.internal RawContent : HTTP/1.1 200 OK Connection: close Accept-Ranges: none Content-Length: 42 Content-Type: text/plain Date: Tue, 20 Sep 2022 15:16:35 GMT Last-Modified: Tue, 20 Sep 2022 14:48:23 GMT Server: EC2ws...
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Moderate: OpenShift Container Platform 4.12.0 bug fix and security update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2022:7399