prestodb/presto · error · PrestoException

NO_CPP_SIDECARS

NO_CPP_SIDECARS

Error message

Expected exactly one coordinator sidecar, but found none

What it means

PluginNodeManager.getSidecarNode returns a coordinator sidecar (native C++ worker attached to the coordinator). When the underlying NodeManager reports no coordinator sidecars, the library throws NO_CPP_SIDECARS because the requested execution target does not exist in this deployment.

Source

Thrown at presto-main-base/src/main/java/com/facebook/presto/nodeManager/PluginNodeManager.java:83

    {
        //Retrieves all active worker nodes, excluding coordinators, resource managers, and catalog servers.
        return nodeManager.getAllNodes().getActiveNodes().stream()
                .filter(node -> !node.isResourceManager() && !node.isCoordinator() && !node.isCatalogServer())
                .collect(toImmutableSet());
    }

    @Override
    public Node getCurrentNode()
    {
        return nodeManager.getCurrentNode();
    }

    @Override
    public Node getSidecarNode()
    {
        Set<InternalNode> coordinatorSidecars = nodeManager.getCoordinatorSidecars();
        if (coordinatorSidecars.isEmpty()) {
            throw new PrestoException(NO_CPP_SIDECARS, "Expected exactly one coordinator sidecar, but found none");
        }
        return coordinatorSidecars.stream()
                .skip(new Random().nextInt(coordinatorSidecars.size()))
                .findFirst()
                .get();
    }

    @Override
    public String getEnvironment()
    {
        return environment;
    }
}

View on GitHub (pinned to 55bb57d202)

Solutions

Verify sidecar processes are launched and configured to register with the coordinator (check sidecar/node config and logs)
Check coordinator /v1/node for registered sidecars; wait for registration or restart the sidecar
Upgrade/pin coordinator and sidecar versions so sidecar registration is enabled; if sidecars are not used, avoid code paths calling getSidecarNode

Example fix

// before
Node sidecar = pluginNodeManager.getSidecarNode(); // throws if none
// after
Set<InternalNode> sidecars = nodeManager.getCoordinatorSidecars();
if (sidecars.isEmpty()) { fallbackToLocalOrThrowWithContext(); }
Node sidecar = sidecars.iterator().next();

Defensive patterns

Strategy: retry

Validate before calling

Set<InternalNode> sidecars = nodeManager.getCoordinatorSidecars();
if (sidecars.isEmpty()) {
    throw new IllegalStateException("No coordinator sidecar registered; check sidecar deployment");
}

Try / catch

try {
    return pluginNodeManager.getSidecarNode();
} catch (PrestoException e) {
    if (e.getErrorCode().getName().equals("NO_CPP_SIDECARS")) {
        // retry after delay (sidecar may still be registering) or fall back to local execution
    }
    throw e;
}

Prevention

Monitor /v1/node on the coordinator for sidecar registration health
Alert when sidecar processes are down on coordinators
Pin coordinator and sidecar versions to compatible releases
Delay sidecar-dependent queries until nodes are registered after startup

When it happens

Trigger: Calling getSidecarNode (directly or via operators that offload to sidecars) on a cluster where no coordinator sidecar is registered — e.g. native sidecars not started or not yet registered with the coordinator.

Common situations: Java-only deployment queried by code expecting C++ sidecars; sidecar process crashed or failed to register; misconfigured node properties (sidecar port/discovery); query run immediately after coordinator start before registration completes.

Related errors

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/88d608f29fe03ba0. Report an issue: GitHub.