Skip to content

A replicated restart_service on Harper core alone ends ERROR instead of restarting this node #3508

Open

Description

@dawsontoth

Part of #641.

Problem

On Harper core alone, a restart_service with replicated: true runs as a job, and the job's first step (restartService in bin/restart.ts) calls server.replication.monitorNodeCAs() to load the cluster's nodes. Core's replication stand-in (server/Server.ts) throws Replication not implemented. there. So the job ends ERROR before it restarts even this node's workers. harper deploy restart=rolling reaches this path when a deploy is not certified (certification: "unavailable"), and the CLI reference documents the ERROR.

The activation path for a certified rolling deploy already treats the stand-in's error as "no peers" (harper#3502): the stand-ins now throw with code: 'REPLICATION_NOT_IMPLEMENTED'.

Done when

A replicated restart_service on core alone restarts this node and completes, treating REPLICATION_NOT_IMPLEMENTED as a cluster with no other nodes, and the CLI reference no longer says such a job ends ERROR.

Review-Deferral: 711b3ce1ee5d8f30 severity=nit category=reliability

🤖 Generated by Claude Code (Claude Opus 5.5); posted via @dawsontoth.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    Fields

    Priority

    P3

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions