NullPointerException from ComputeEngineComputerLauncher.launch

[jglick]

Jenkins and plugins versions report

Environment

Jenkins 2.319.3 (CloudBees CI variant), google-compute-engine 4.3.8

What Operating System are you using (both controller, and any agents involved in the problem)?

Linux

Reproduction steps

Unknown exactly. Jenkins was upgraded while a Pipeline build was running that used a GCE agent. After restart, the build tried to resume but failed, perhaps because the agent VM had been killed. While normally there is a 5m timeout waiting for the agent to reconnect before Pipeline infrastructure gives up and fails (pending jenkinsci/workflow-durable-task-step-plugin#180), in this case the build seemed to keep on trying to resume longer than that. I went to https://jenkins/computer/gce-cloud-…/log and saw an error.

Expected Results

No exceptions, perhaps just an error message saying that the VM could not be relaunched.

Actual Results

java.lang.NullPointerException
	at com.google.jenkins.plugins.computeengine.ComputeEngineComputerLauncher.launch(ComputeEngineComputerLauncher.java:157)
	at hudson.slaves.SlaveComputer.lambda$_connect$0(SlaveComputer.java:293)
	at …

which corresponds to

google-compute-engine-plugin/src/main/java/com/google/jenkins/plugins/computeengine/ComputeEngineComputerLauncher.java

Lines 154 to 158 in 9dfd5b7

	Operation operation =
	cloud
	.getClient()
	.waitForOperationCompletion(
	cloud.getProjectId(), insertOperationId, zone, node.getLaunchTimeoutMillis());

suggesting a failure in

google-compute-engine-plugin/src/main/java/com/google/jenkins/plugins/computeengine/ComputeEngineCloud.java

Lines 188 to 197 in 9dfd5b7

	private ComputeClient createClient() {
	try {
	ClientFactory clientFactory = ClientUtil.getClientFactory(Jenkins.get(), credentialsId);
	return clientFactory.computeClient();
	} catch (IOException e) {
	log.log(Level.SEVERE, "Exception when creating GCE client", e);
	// TODO: https://github.com/jenkinsci/google-compute-engine-plugin/issues/62
	return null;
	}
	}

or #62. Looking for that message in log files, I see

SEVERE	c.g.j.p.c.ComputeEngineCloud#createClient: Exception when creating GCE client
hudson.AbortException: Failed to initialize HTTP transport: hudson.AbortException: Could not retrieve credentials: …
	at com.google.jenkins.plugins.computeengine.client.ClientUtil.getClientFactory(ClientUtil.java:56)
	at com.google.jenkins.plugins.computeengine.client.ClientUtil.getClientFactory(ClientUtil.java:71)
	at com.google.jenkins.plugins.computeengine.ComputeEngineCloud.createClient(ComputeEngineCloud.java:190)
	at com.google.jenkins.plugins.computeengine.ComputeEngineCloud.getClient(ComputeEngineCloud.java:208)
	at com.google.jenkins.plugins.computeengine.ComputeEngineComputerLauncher.launch(ComputeEngineComputerLauncher.java:156)
	at hudson.slaves.SlaveComputer.lambda$_connect$0(SlaveComputer.java:293)
	at …

pointing to

google-compute-engine-plugin/src/main/java/com/google/jenkins/plugins/computeengine/client/ClientUtil.java

Lines 50 to 57 in 9dfd5b7

	try {
	GoogleRobotCredentials robotCreds =
	getRobotCredentials(itemGroup, domainRequirements, credentialsId);
	Credential googleCredential = getGoogleCredential(robotCreds);
	clientFactory = new ClientFactory(transport, googleCredential, APPLICATION_NAME);
	} catch (IOException | GeneralSecurityException ex) {
	throw new AbortException(Messages.ClientFactory_FailedToInitializeHTTPTransport(ex));
	}

Note that the AbortException here loses track of the nested IOException.

Anything else?

No response