What causes "Failed to reap orphaned chroma-mcp descendant" in thedotmack/claude-mem?

Check the logged errno: EPERM means run the cleanup under the same user/with sufficient privileges.

Failed to reap orphaned chroma-mcp descendant

Error message

Failed to reap orphaned chroma-mcp descendant

What it means

During cleanup the manager walks orphaned chroma-mcp descendant processes and SIGKILLs them. If process.kill fails with an errno other than ESRCH (process already gone), this warning is logged; the orphan may still be running.

Solutions

Check the logged errno: EPERM means run the cleanup under the same user/with sufficient privileges.
Verify the pid still belongs to a chroma-mcp descendant and kill it manually (kill -9 <pid>).
In containers, ensure the reaper shares the PID namespace and has CAP_KILL if needed.
If benign and recurring, confirm the process exits on its own and the warning is ignorable.

Example fix

// before
if (code !== 'ESRCH') { logger.warn(...); }
// after: handle EPERM explicitly with guidance
if (code === 'EPERM') {
  logger.warn('CHROMA_MCP', `Cannot kill orphan ${pid}: permission denied; kill manually`);
} else if (code !== 'ESRCH') {
  logger.warn('CHROMA_MCP', 'Failed to reap orphaned chroma-mcp descendant', { pid, code });
}

Defensive patterns

Strategy: try-catch

Validate before calling

// check you can signal the pid first
try {
  process.kill(pid, 0); // existence/permission probe
} catch (e) {
  console.log(`pid ${pid} not signallable: ${e.code}`);
}

Try / catch

try { process.kill(pid, 'SIGKILL'); }
catch (e) {
  const code = (e as NodeJS.ErrnoException).code;
  if (code !== 'ESRCH') console.warn(`reap failed pid=${pid} code=${code}`);
}

Prevention

Run the app and its cleanup under the same user
Avoid reaping pids your process did not spawn
In containers, keep children in the same PID namespace

When it happens

Trigger: SIGKILL on a descendant pid raises EPERM (owned by another user), EINVAL, or another errno. Only ESRCH (no such process) is silently tolerated.

Common situations: Reaper running as a different user than the process that spawned chroma-mcp, containerized setups restricting signals between processes, or pid reuse races.

Understand the failure class

Background: Permission denied / not authorized / 403 Forbidden: access-control rejections when the caller lacks the required role, grant, or ownership — this error's family across 18 libraries.

Related errors

AI-assisted analysis of thedotmack/claude-mem@d8bc9755e7 (2026-09-17). Data as JSON: /api/errors/4840fa7f49f8db73. Report an issue: GitHub.

Appendix: source

Thrown at src/services/sync/ChromaMcpManager.ts:1178

    const seen = new Set<number>();

    for (const { pid, startToken } of entries) {
      if (seen.has(pid)) continue;
      seen.add(pid);
      if (!isPidAlive(pid)) continue;

      if (!isSameProcess(pid, startToken)) {
        logger.debug('CHROMA_MCP', 'Skipping reap: PID was recycled since the snapshot', { pid });
        continue;
      }

      try {
        process.kill(pid, 'SIGKILL');
        logger.debug('CHROMA_MCP', 'Reaped orphaned chroma-mcp descendant', { pid });
      } catch (error) {
        const code = (error as NodeJS.ErrnoException).code;
        if (code !== 'ESRCH') {
          logger.warn('CHROMA_MCP', 'Failed to reap orphaned chroma-mcp descendant', {
            pid,
            code,
          });
        }
      }
    }
  }

  private async disposeActivePrewarm(): Promise<void> {
    const prewarmChild = this.activePrewarmChild;
    const tracked = this.activePrewarmTracked;
    if (!prewarmChild) {
      return;
    }
    if (this.activePrewarmChild === prewarmChild) {
      this.activePrewarmChild = null;
      this.activePrewarmTracked = null;
    }

View on GitHub (pinned to d8bc9755e7)