Ubuntu 24.04 LTS Cloud Image ova error during deployment

[Aditya Irawan]

Ubuntu 24.04 LTS Cloud Image daily (20240416)
File: noble-server-cloudimg-amd64.ova
Error: OVF's rasd item with instance ID=4 has invalid serial port config.
Deploy to: VMware Cloud Director 10.4.2.22463311, VMware ESXi, 7.0.3, 20842708

Detail debug:
com.vmware.vcloud.api.presentation.service.BadRequestException: OVF's rasd item with instance ID=4 has invalid serial port config.
 at com.vmware.vcloud.fabric.storage.copy.impl.OvfTransferManagerImpl.validateSerialPortConfig(OvfTransferManagerImpl.java:737)
 at com.vmware.ssdc.backend.services.impl.VAppUploadManagerImpl.handleUploadOvfDescriptor(VAppUploadManagerImpl.java:2180)
 at com.vmware.ssdc.backend.services.impl.VAppUploadManagerImpl.handleUploadBody(VAppUploadManagerImpl.java:2060)
 at com.vmware.ssdc.backend.services.impl.VAppUploadManagerImpl.handleUpload(VAppUploadManagerImpl.java:1957)
 at com.vmware.ssdc.backend.services.impl.VAppUploadManagerImpl.executeTask(VAppUploadManagerImpl.java:5285)
 at com.vmware.vcloud.backendbase.management.system.TaskActivity$ExecutePhase$1.doInSecurityContext(TaskActivity.java:828)
 at com.vmware.vcloud.backendbase.management.system.TaskActivity$ExecutePhase$1.doInSecurityContext(TaskActivity.java:823)
 at com.vmware.vcloud.backendbase.management.system.SecurityContextTemplate.executeForOrgAndUser(SecurityContextTemplate.java:49)
 at com.vmware.vcloud.backendbase.management.system.TaskActivity$ExecutePhase.execute(TaskActivity.java:830)
 at com.vmware.vcloud.backendbase.management.system.TaskActivity$ExecutePhase.invokeInner(TaskActivity.java:726)
 at com.vmware.vcloud.backendbase.management.system.TaskActivity$TaskActivityBasePhase.invoke(TaskActivity.java:342)
 at com.vmware.vcloud.activity.executors.ActivityRunner.runPhase(ActivityRunner.java:175)
 at com.vmware.vcloud.activity.executors.ActivityRunner.run(ActivityRunner.java:112)
 at java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
 at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
 at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
 at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
 at java.base/java.lang.Thread.run(Thread.java:829)

It appears that .ovf file contains serial port configuration.
      <Item ovf:required="false">
        <rasd:AutomaticAllocation>false</rasd:AutomaticAllocation>
        <rasd:ElementName>serial0</rasd:ElementName>
        <rasd:InstanceID>4</rasd:InstanceID>
        <rasd:ResourceSubType>vmware.serialport.device</rasd:ResourceSubType>
        <rasd:ResourceType>21</rasd:ResourceType>
        <vmw:Config ovf:required="false" vmw:key="yieldOnPoll" vmw:value="false" />
      </Item>

OVF file shouldn't include serial port: https://kb.vmware.com/s/article/2128084

I was able to deploy the cloud image after removing serial port configuration from the ovf file.

[John Chittum (jchittum)]

1. what versions of VMware ESXi is this applicable?
2. is this reproducible in other environments, such as Virtualbox?
3. is this error true of all ubuntu images for those versions then? This was added in https://bugs.launchpad.net/ubuntu/+source/livecd-rootfs/+bug/1895104 at the request of vmware
4. I have a worry that removing serial port may cause issues in other deployment areas, such as Virtualbox, and possibly other versions of ESXi. Could a full test plan be provided to ensure compatibility?

This is also tightly coupled to this ticket, in which i requested a meeting and documentation, and it hasn't occurred yet: https://bugs.launchpad.net/ubuntu/+bug/1898871

NOTE: this is after Release Freeze so there is a change this does _not_ get fixed prior to the 24.04 release next week. If we cannot work out a fix, we will release note this issue and work on releasing as soon as possible

[Aditya Irawan (air-awan)]

1. I was testing this on VMware ESXi, 7.0.3, 20842708 (that is ESXi 7 Update 3i).
3. Since you mentioned this, I did further investigation.
I had deployed Ubuntu 22.04 (Jammy Jellyfish) [20230302] successfully before, and you're right, its ovf file contained the exact same serial port configuration. But when I tried to deploy it again just now, it failed.
I suspect this has got to do with the Cloud Director, because the deployment process has not reached vSphere yet. Also, between the last successful cloud image deployment and now, we have not do any changes to the ESXi version, but did an upgrade to the Cloud Director 10.4.2.22463311.
So I believe this has got nothing to do with cloud images as it worked before and will try to raise an issue with VMware regarding this error.

[John Chittum (jchittum)]

Thank you for checking Aditya. I'll leave this open till you get some guidance from VMware regarding any changes in Cloud Director that may be affecting serial console connections.

I've also added this to our 24.04 release notes (to be published soon for the clouds). Thank you a bunch for trying out 24.04 so that we can help others if they see this as well!

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[John Chittum (jchittum)]

I've set this to incomplete for now. @air-awan, if you have time to give an update related to vmware environments, let me know.

i double checked, and the addition of the serial port was requested by the vmware team, so if it's breaking things now due to hypervisor level issues, I'll have to sync with vmware to figure out the appropriate fix.