quarkusio/quarkus · warning · RuntimeException

Error closing file:

Error message

Error closing file: 

What it means

While finishing deployment, deploy() closes the reader over the generated manifest; if closing throws IOException it is rethrown as RuntimeException 'Error closing file: <path>'. This indicates an I/O problem releasing the manifest file handle rather than a deployment logic problem.

Source

Thrown at extensions/kubernetes/vanilla/deployment/src/main/java/io/quarkus/kubernetes/deployment/KubernetesDeployer.java:236

                deployResource(deploymentTarget, client, i, optionalResourceDefinitions);
                log.info("Applied: " + i.getKind() + " " + i.getMetadata().getName() + ".");
            });

            printExposeInformation(client, list, openshiftConfig, applicationInfo);

            HasMetadata m = list.getItems().stream()
                    .filter(r -> deploymentTarget.getDeploymentResourceKind().matches(r))
                    .findFirst().orElseThrow(() -> new IllegalStateException(
                            "No " + deploymentTarget.getDeploymentResourceKind() + " found under: "
                                    + manifest.getAbsolutePath()));
            return new DeploymentResultBuildItem(m.getMetadata().getName(), m.getMetadata().getLabels());
        } catch (FileNotFoundException e) {
            throw new IllegalStateException("Can't find generated kubernetes manifest: " + manifest.getAbsolutePath());
        } catch (KubernetesClientException e) {
            KubernetesClientErrorHandler.handle(e);
            throw e;
        } catch (IOException e) {
            throw new RuntimeException("Error closing file: " + manifest.getAbsolutePath());
        }

    }

    private void deployResource(DeploymentTargetEntry deploymentTarget, KubernetesClient client, HasMetadata metadata,
            List<KubernetesOptionalResourceDefinitionBuildItem> optionalResourceDefinitions) {
        var r = findResource(client, metadata);
        Optional<HasMetadata> existing = Optional.ofNullable(client.resource(metadata).get());
        if (shouldDeleteExisting(deploymentTarget, metadata)) {
            deleteResource(metadata, r);
        }

        try {
            switch (deploymentTarget.getDeployStrategy()) {
                case Create:
                    r.create();
                    break;
                case Replace:

View on GitHub (pinned to e1c734241f)

Solutions

1. Retry the build; transient filesystem contention is the usual cause
2. Ensure no other process deletes/replaces target/kubernetes manifests while the build runs (parallel builds, IDE cleaners)
3. Check disk space and filesystem health on the machine running the build

Example fix

// before
mvn package -Dquarkus.kubernetes.deploy=true &  mvn clean ...
// after (avoid concurrent access to target/)
mvn clean package -Dquarkus.kubernetes.deploy=true

if (!java.nio.file.Files.isReadable(manifestPath)) throw new IllegalStateException("manifest not readable: " + manifestPath);

try { deploy(manifest); } catch (RuntimeException e) {
    if (e.getMessage().startsWith("Error closing file")) {
        LOG.warnf("I/O issue closing %s — check disk/filesystem and retry", manifestPath);
    }
    throw e;
}

When it happens

Trigger: The try-with-resources/finally close of the manifest file reader in deploy() throws IOException — e.g. disk/descriptor issues while the file handle is released.

Common situations: Filesystem issues (full disk, NFS stalls, container volume unmounted) during builds; file being modified/removed concurrently by another process while being read; rare JDK-level I/O failures.

Related errors

• Can't find generated kubernetes manifest:
• Unable to copy json config file from ${jsonPath} to ${thinJa
• Unable to determine groupId and artifactId of the jar that c
• Build:%s is no longer present!
• Build:%s has no status!

AI-assisted analysis of quarkusio/quarkus@e1c734241f (2026-09-05). Data as JSON: /api/errors/2ccc82da618a9f8c. Report an issue: GitHub.