live wire
▸JAVA · Quarkus 4.0.0.Beta1 moves to Java 21, adds HTTP/3 and starts extension migration (Oct. 1)Quarkus▸SECURITY · X41 shows shared /dev/shm can turn Envoy hot restart into cross-container lateral movementX41 D-Sec▸DATA · AWS and Red Hat map Confluent Platform on ROSA with HCP, CFK and OpenShift security controlsAWS IBM & Red Hat▸API · Red Hat resolves intermittent 3scale API Manager latencyRed Hat Status▸AI · IBM shows Maximo workflows exposed as approval-gated MCP tools on OpenShiftIBM Community▸AI · vLLM adds day-zero NVIDIA Vera Rubin support and reports 7.8× per-GPU throughputvLLM▸INTEGRATION · Apache Camel 4.23 makes Kamelets visible to AI tooling and validationApache Camel▸SECURITY · OpenShift 4.14.75 fixes five CVEs, including two SQLite code-execution flawsRed Hat Customer Portal▸SUPPLY CHAIN · Red Hat maps CRA-ready open source practices as EU reporting rules take effectRed Hat Blog▸AI · Red Hat AI Inference on IBM Cloud adds an OpenAI-compatible Embeddings APIIBM Cloud▸API · Red Hat investigates degraded 3scale API Management SaaS APIsRed Hat Status▸PLATFORM · Red Hat and Cloudera validate a 100-VM analytics stack on OpenShift VirtualizationRed Hat Blog▸DEVELOPER HUB · Red Hat maps a four-zone, quota-aware Dev Spaces architectureRed Hat Developer▸INTEGRATION · Camel 4.23 teaches agent tools to discover and validate KameletsApache Camel▸JAVA · Quarkus 4.0.0.Beta1 moves to Java 21, adds HTTP/3 and starts extension migration (Oct. 1)Quarkus▸SECURITY · X41 shows shared /dev/shm can turn Envoy hot restart into cross-container lateral movementX41 D-Sec▸DATA · AWS and Red Hat map Confluent Platform on ROSA with HCP, CFK and OpenShift security controlsAWS IBM & Red Hat▸API · Red Hat resolves intermittent 3scale API Manager latencyRed Hat Status▸AI · IBM shows Maximo workflows exposed as approval-gated MCP tools on OpenShiftIBM Community▸AI · vLLM adds day-zero NVIDIA Vera Rubin support and reports 7.8× per-GPU throughputvLLM▸INTEGRATION · Apache Camel 4.23 makes Kamelets visible to AI tooling and validationApache Camel▸SECURITY · OpenShift 4.14.75 fixes five CVEs, including two SQLite code-execution flawsRed Hat Customer Portal▸SUPPLY CHAIN · Red Hat maps CRA-ready open source practices as EU reporting rules take effectRed Hat Blog▸AI · Red Hat AI Inference on IBM Cloud adds an OpenAI-compatible Embeddings APIIBM Cloud▸API · Red Hat investigates degraded 3scale API Management SaaS APIsRed Hat Status▸PLATFORM · Red Hat and Cloudera validate a 100-VM analytics stack on OpenShift VirtualizationRed Hat Blog▸DEVELOPER HUB · Red Hat maps a four-zone, quota-aware Dev Spaces architectureRed Hat Developer▸INTEGRATION · Camel 4.23 teaches agent tools to discover and validate KameletsApache Camel
upstreambeat.ai
guideJAVA

Standalone S2I brings Red Hat OpenJDK container builds into the local development loop

Red Hat’s walkthrough shows how teams can reproduce OpenShift-style Java image assembly locally, inspect it and carry configuration forward.

Local and cluster image builds compared.
Side by side: what changed
By The News Desk· Sep 23, 2026the quick take — two AI hosts go live when you do

Red Hat has published a hands-on guide for building and testing Red Hat OpenJDK application images locally with the standalone Source-to-Image (S2I) command-line tool before deploying them to OpenShift. The workflow uses the same self-assembling builder-image model familiar from OpenShift builds, but it runs outside the cluster and does not require Shipwright or Jenkins. The Red Hat Developer walkthrough covers current OpenShift 4 and Red Hat build of OpenJDK releases.

What the local loop provides

S2I injects application source into a builder image, executes that image’s assembly scripts and commits the result as a runnable application image. Red Hat’s examples build Spring Boot and WAR-based applications from either a local checkout or a remote Git repository using UBI-based OpenJDK 11 and 17 builder images. The resulting image can then be run locally and later moved into the deployment pipeline. The guide includes complete command output for checkout, Maven packaging, artifact placement and image commit.

That visibility is the practical advantage. Running s2i build with a higher log level exposes which builder image, scripts, environment and Maven command actually produced the container. Developers can diagnose assembly behavior before consuming OpenShift build capacity, while platform teams can point both local and cluster workflows at controlled builder-image tags. Red Hat also documents --as-dockerfile, pull-policy, context-directory and script-location options.

Configuration travels with the source

The walkthrough uses .s2i/environment and --environment-file to pass JVM and Maven settings into the build. Examples replace garbage-collector and memory defaults through MAVEN_OPTS, supply a custom Maven settings file with MAVEN_SETTINGS_XML, and define repeatable source, builder and output-image parameters. That makes the local image build more than a one-off shell command: the application repository can carry its expected assembly inputs. The same guide contrasts this with OpenShift BuildConfig source and binary-build examples.

Where teams should be careful

The article’s command transcripts mix docker, privileged execution and exact dated builder-image tags. Teams should adapt the container engine and image references to their supported workstation policy rather than copying those details blindly. They should also keep credentials and private Maven repository settings out of committed environment files, even though the mechanism can pass a custom settings path into the build.

For platform teams, the useful pattern is to publish an approved OpenJDK builder image and a small, versioned S2I configuration that developers can run locally and CI can reproduce. The cluster remains the deployment authority, but the assembly contract becomes testable earlier. Red Hat’s documented workflow provides the local-to-OpenShift bridge and the log points needed to verify it.

Filed by The News Desk. Corrections: desk@upstreambeat.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.