live wire
▸PLATFORM · Clang 23 bytecode interpreter cuts allocations and expands constant-expression supportRed Hat Developer▸PLATFORM · RHEL becomes first-party on OCI; OpenShift subscriptions reach Oracle MarketplaceRed Hat▸AI · vLLM-Omni merges native MOSS-Image serving with parallelism and offloadvLLM▸AI · vLLM adds System One endpoint for Decision 3.0 multimodal routing models (Oct. 15)vLLM▸SECURITY · Red Hat rates OpenShift AI 3.4.5 update Critical; advisory lists hundreds of CVEsRed Hat Product Security▸SECURITY · Red Hat’s first Security Signals report shifts focus from CVE counts to risk paths (Oct. 15)Red Hat▸PLATFORM · IBM adds non-destructive FlexVolume-to-CSI migration tool for OpenShift object storageIBM Cloud▸vLLM tests agent-built Kimi K3 decode path at up to 1.9× speedup; PR remains openvLLM▸AI · LeaderWorkerSet 1.0 changes rollout defaults for disaggregated AI workloadsKubernetes SIGs▸SECURITY · Red Hat flags active exploitation of kernel flaw affecting OpenShift CoreOS; no fix listedRed Hat Product Security▸AI · Red Hat sets Dec. 9 showcase for AI 3.6 agent governance and model servicesRed Hat▸vLLM 0.32.0 drops Python 3.10, restricts CORS defaults and adds security fixesvLLM▸SECURITY · Red Hat tracks Moderate Hono auth bypass across OpenShift AI, Dev Spaces and AnsibleRed Hat Product Security▸SECURITY · Red Hat Hardened Images fixes Moderate libpng denial-of-service flawRed Hat▸PLATFORM · Clang 23 bytecode interpreter cuts allocations and expands constant-expression supportRed Hat Developer▸PLATFORM · RHEL becomes first-party on OCI; OpenShift subscriptions reach Oracle MarketplaceRed Hat▸AI · vLLM-Omni merges native MOSS-Image serving with parallelism and offloadvLLM▸AI · vLLM adds System One endpoint for Decision 3.0 multimodal routing models (Oct. 15)vLLM▸SECURITY · Red Hat rates OpenShift AI 3.4.5 update Critical; advisory lists hundreds of CVEsRed Hat Product Security▸SECURITY · Red Hat’s first Security Signals report shifts focus from CVE counts to risk paths (Oct. 15)Red Hat▸PLATFORM · IBM adds non-destructive FlexVolume-to-CSI migration tool for OpenShift object storageIBM Cloud▸vLLM tests agent-built Kimi K3 decode path at up to 1.9× speedup; PR remains openvLLM▸AI · LeaderWorkerSet 1.0 changes rollout defaults for disaggregated AI workloadsKubernetes SIGs▸SECURITY · Red Hat flags active exploitation of kernel flaw affecting OpenShift CoreOS; no fix listedRed Hat Product Security▸AI · Red Hat sets Dec. 9 showcase for AI 3.6 agent governance and model servicesRed Hat▸vLLM 0.32.0 drops Python 3.10, restricts CORS defaults and adds security fixesvLLM▸SECURITY · Red Hat tracks Moderate Hono auth bypass across OpenShift AI, Dev Spaces and AnsibleRed Hat Product Security▸SECURITY · Red Hat Hardened Images fixes Moderate libpng denial-of-service flawRed Hat
upstreambeat.ai
newsJAVA

Quarkus MCP Server 2.0 makes stateless MCP deployable without a flag day

The extension accepts the July 2026 stateless protocol and legacy session-based clients on one endpoint, giving platform teams a staged migration path.

Legacy and stateless MCP requests share one Quarkus endpoint during migration.
AI-generated illustration
By The News Desk· Sep 26, 2026the quick take — two AI hosts go live when you do

Quarkus has added support for the Model Context Protocol’s stateless transport model without forcing existing clients to migrate at the same time. The 2.0.x line of Quarkus MCP Server handles the July 28, 2026 protocol revision and older session-based versions on the same endpoint, according to the project’s September 21 engineering post.

That compatibility detail matters more than the version number. Under the earlier design, a client initialized a session, received an identifier and had to keep talking to the same server instance—or rely on sticky routing or a shared session store. The current MCP specification instead defines stateless, self-contained JSON-RPC requests with capability negotiation performed per request.

A migration path rather than a cutover

Quarkus MCP Server inspects the protocol version carried in the request metadata and HTTP headers. Requests using the July 2026 revision are handled through a transient connection for the life of the call; older requests fall back to the established stateful path. Both modes remain available at the same URL.

For application teams, that means clients can move individually rather than through a coordinated cutover. Quarkus LangChain4j clients can discover server capabilities and prefer the new revision automatically, while operators can pin either the stateless or legacy protocol during a migration.

The operational payoff is straightforward: any available server instance can process a self-contained request. That removes protocol-level pressure for session affinity when horizontally scaling an MCP service behind a load balancer.

Server-to-client calls change shape

Stateless operation also changes features that previously depended on a persistent channel. Sampling, elicitation and roots can no longer assume that a server will call back across an open session. The new protocol represents those interactions as multi-round-trip requests: a server returns an input_required result, and the client retries the original request with the requested data attached.

Quarkus exposes both paths through its APIs so tools can detect whether server-initiated requests are supported and use the appropriate flow. Subscriptions similarly move from the older resources/subscribe approach to a filtered subscriptions/listen stream for stateless clients.

What teams need to do

Existing Quarkus MCP Server users can adopt the 2.0.x release and begin accepting newer stateless clients without rewriting their tool methods. Code that uses sampling, elicitation or roots needs closer review because it may have to handle both callback and retry-based interactions.

The project also says it runs the official MCP conformance suite in continuous integration. Support for the revision’s extension mechanism is planned for Quarkus MCP Server 2.1.0 rather than included in the current line.

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.