prestodb/presto · critical · TException

Failed connecting to Hive metastore: ${addresses}

Error message

Failed connecting to Hive metastore: ${addresses}

What it means

StaticHiveCluster.createMetastoreClient loops over every configured metastore URI and attempts a Thrift connection; if all attempts fail with a TException, it throws this aggregate TException with the list of addresses and the last underlying exception as its cause. It means Presto could not open a working Thrift session to any Hive metastore host in the static cluster.

Source

Thrown at presto-hive-metastore/src/main/java/com/facebook/presto/hive/metastore/thrift/StaticHiveCluster.java:89

        if (metastoreLoadBalancingEnabled) {
            Collections.shuffle(metastores);
        }

        TException lastException = null;
        for (HostAndPort metastore : metastores) {
            try {
                HiveMetastoreClient client = clientFactory.create(metastore, token);

                if (!isNullOrEmpty(metastoreUsername)) {
                    client.setUGI(metastoreUsername);
                }
                return client;
            }
            catch (TException e) {
                lastException = e;
            }
        }
        throw new TException("Failed connecting to Hive metastore: " + addresses, lastException);
    }

    private static URI checkMetastoreUri(URI uri)
    {
        requireNonNull(uri, "metastoreUri is null");
        String scheme = uri.getScheme();
        checkArgument(!isNullOrEmpty(scheme), "metastoreUri scheme is missing: %s", uri);
        checkArgument(scheme.equals("thrift"), "metastoreUri scheme must be thrift: %s", uri);
        checkArgument(uri.getHost() != null, "metastoreUri host is missing: %s", uri);
        checkArgument(uri.getPort() != -1, "metastoreUri port is missing: %s", uri);
        return uri;
    }
}

View on GitHub (pinned to 55bb57d202)

Solutions

  1. Verify the metastore is up and listening: run `nc -vz <host> 9083` from the Presto coordinator and start hive-metastore service if down
  2. Check hive.metastore.thrift.uris values for typos in host/port/scheme; confirm the port matches hive.metastore.uris on the metastore side
  3. Inspect the `cause` (lastException) of this exception in the coordinator log — it holds the real per-host failure (refused, timeout, SSL, SASL)
  4. Open firewall/security groups for the metastore port and confirm DNS resolves the configured hostnames
  5. If metastore is healthy but slow, raise hive.metastore.thrift.client.connect-timeout / read-timeout and metastore refresh settings

Example fix

// before
hive.metastore=thrift
hive.metastore.thrift.uris=thrift://meta-internal:9083
// after
hive.metastore=thrift
hive.metastore.thrift.uris=thrift://metastore-host.example.com:9083
# verified with: nc -vz metastore-host.example.com 9083
Defensive patterns

Strategy: try-catch

Validate before calling

// Pre-flight: verify at least one metastore URI is reachable before initializing the catalog
List<String> uris = Arrays.asList("thrift://metastore-host.example.com:9083");
for (String uri : uris) {
    try {
        URI u = new URI(uri);
        try (Socket s = new Socket()) {
            s.connect(new InetSocketAddress(u.getHost(), u.getPort()), 3000); // fails fast if refused
        }
    } catch (Exception e) {
        throw new IllegalStateException("Metastore unreachable: " + uri, e);
    }
}

Type guard

public static boolean isConnectionFailure(Throwable t) {
    Throwable root = t;
    while (root.getCause() != null) root = root.getCause();
    return root instanceof ConnectException || root instanceof UnknownHostException
        || root instanceof SocketTimeoutException || root instanceof TTransportException;
}

Try / catch

try {
    HiveMetastoreClient client = staticHiveCluster.createMetastoreClient(metastoreContext);
} catch (TException e) {
    if (isConnectionFailure(e)) {
        log.error("No metastore host reachable; cause={}", rootCause(e).getMessage(), e);
        throw new RetriableException("Metastore cluster unreachable: " + e.getMessage(), e);
    }
    throw e;
}

Prevention

When it happens

Trigger: Any Presto query or catalog operation requiring metastore access when none of the URIs in hive.metastore (or the hive.metastore.thrift.uris list) accept a connection: connection refused, TLS handshake failure, auth failure, or timeout on each host.

Common situations: Metastore service down or never started; wrong hive.metastore.thrift.uris port (default 9083); firewall/security-group blocking; kerberos or SSL misconfiguration on the client; metastore host DNS not resolving inside containers; metastore overloaded and dropping connections.

Understand the failure class

Background: ECONNREFUSED and "connection refused" / "could not connect to server" errors: what they mean and how to fix them — this error's family across 44 libraries.

Related errors

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/a15efafa71020620. Report an issue: GitHub.