mongodb/node-mongodb-native · error · MongoRuntimeError

Unreachable. If you are seeing this error, please file a tic

Error message

Unreachable. If you are seeing this error, please file a ticket on the NODE driver project on Jira

What it means

Thrown by the @internal CSOTTimeoutContext.connectionCheckoutTimeout getter when _serverSelectionTimeout is not an object (null/undefined) at the time the checkout timeout is requested. The code's invariant is that server selection runs first and produces a timeout object that the checkout reuses; reaching this branch means that invariant was violated. It is a MongoRuntimeError explicitly labeled unreachable and asking for a Jira ticket.

Source

Thrown at src/timeout.ts:272

        } else {
          this._serverSelectionTimeout = null;
        }
      }
    }

    return this._serverSelectionTimeout;
  }

  get connectionCheckoutTimeout(): Timeout | null {
    if (
      typeof this._connectionCheckoutTimeout !== 'object' ||
      this._connectionCheckoutTimeout?.cleared
    ) {
      if (typeof this._serverSelectionTimeout === 'object') {
        // null or Timeout
        this._connectionCheckoutTimeout = this._serverSelectionTimeout;
      } else {
        throw new MongoRuntimeError(
          'Unreachable. If you are seeing this error, please file a ticket on the NODE driver project on Jira'
        );
      }
    }
    return this._connectionCheckoutTimeout;
  }

  get timeoutForSocketWrite(): Timeout | null {
    const { remainingTimeMS } = this;
    if (!Number.isFinite(remainingTimeMS)) return null;
    if (remainingTimeMS > 0) return Timeout.expires(remainingTimeMS);
    return Timeout.reject(new MongoOperationTimeoutError('Timed out before socket write'));
  }

  get timeoutForSocketRead(): Timeout | null {
    const { remainingTimeMS } = this;
    if (!Number.isFinite(remainingTimeMS)) return null;
    if (remainingTimeMS > 0) return Timeout.expires(remainingTimeMS);

View on GitHub (pinned to 3366c21a63)

Solutions

1. Report it as a driver bug via the NODE project on Jira with the driver version, topology, timeoutMS, and a repro.
2. As a workaround, disable CSOT (drop timeoutMS) for the affected operation to avoid the CSOT code path.
3. Upgrade to the latest driver release where the ordering may have been fixed.

Example fix

// No user-side code fix - this is a driver invariant violation.
// Workaround: avoid the CSOT path.

// before
await coll.find({}, { timeoutMS: 5000 }).toArray(); // triggers CSOT path

// after (workaround)
await coll.find({}).toArray(); // legacy timeout path

// no user-side validation prevents this internal invariant violation;
// work around by avoiding the CSOT path:
await coll.find({}).toArray(); // no timeoutMS -> legacy timeout path

try {
  await coll.find({}, { timeoutMS: 5000 }).toArray();
} catch (e) {
  if (e instanceof MongoRuntimeError && /Unreachable/.test(e.message)) {
    // fall back to no-CSOT path and file a driver bug
    await coll.find({}).toArray();
  } else throw e;
}

When it happens

Trigger: Driver-internal race or ordering bug where connectionCheckoutTimeout is accessed before serverSelectionTimeout has been computed, or after it was cleared to null. Not user-reachable through normal API use.

Common situations: Unusual CSOT + load-balancer or transaction pinning sequences; regressions after timeout-context refactors; concurrent access to a single TimeoutContext from multiple operations.

Related errors

• Unrecognized options
• Cannot create a Timeout with a negative duration
• Timed out
• KMS request timed out
• Unexpected HostAddress ${JSON.stringify(hostAddress)}

AI-assisted analysis of mongodb/node-mongodb-native@3366c21a63 (2026-08-04). Data as JSON: /data/errors/3a9aa3f571049f3b.json. Report an issue: GitHub.