What causes "Found faulty validators" in linera-io/linera-protocol?

Use the printed 'Faulty validators:' list to identify which node misbehaved, then inspect its logs

Found faulty validators

Error message

Found faulty validators

What it means

After querying validators, this run() prints every validator classified as faulty — name, address, and its accumulated error count — and then bails. 'Faulty' means the validator was reached but its behavior/answers during the query session indicated problems (errors while processing requests), as distinct from merely unreachable validators. The printed list is the diagnostic; the bail makes the failure visible in the exit code.

Source

Thrown at linera-service/src/cli/validator.rs:698

        println!("Local Node:");
        local_results.print(None, None, None, None);

        // Print validator results (only differences from local node)
        for (name, address, votes, results) in &validator_results {
            results.print(
                Some(name),
                Some(address),
                Some(*votes),
                Some(&local_results),
            );
        }

        if !faulty_validators.is_empty() {
            println!("\nFaulty validators:");
            for ((name, address), errors) in faulty_validators {
                println!("  {} at {}: {} error(s)", name, address, errors.len());
            }
            anyhow::bail!("Found faulty validators");
        }

        Ok(())
    }
}

impl Query {
    async fn run(
        &self,
        context: &ClientContext<impl linera_core::Environment>,
    ) -> anyhow::Result<()> {
        let node = context.make_node_provider().make_node(&self.address)?;
        let chain_id = self.chain_id.unwrap_or_else(|| context.default_chain());
        println!("Querying validator about chain {chain_id}.\n");

        let results = context
            .query_validator(&self.address, &node, chain_id, self.public_key.as_ref())
            .await;

View on GitHub (pinned to 6c226ddcb3)

Solutions

Use the printed 'Faulty validators:' list to identify which node misbehaved, then inspect its logs
Restart the faulty validator(s) and re-run the query
For local nets with corrupted state, tear down and re-create: `linera net down` then `linera net up`
Verify client and validator binary versions match (version skew often looks like faults)

Defensive patterns

Strategy: try-catch

Try / catch

match run(&mut net).await {
    Err(e) if e.to_string().contains("Found faulty validators") => {
        // the faulty list (name/address/error count) was printed to stdout;
        // capture and parse it to decide: restart vs recreate the net
        handle_faulty_validators(&captured_output);
    }
    result => result?,
}

Prevention

Run the query command with stdout captured so the printed faulty list is machine-parseable on failure
Keep client and validator binaries on the same version — skew frequently masquerades as faults
For fault-injection experiments, expect this bail and assert on it rather than treating it as spurious

When it happens

Trigger: Running the validator inspection command when `faulty_validators` is non-empty: validators that produced errors during the query session (invalid responses, protocol-level errors, corrupted state).

Common situations: Fault-injection experiments on local test nets; a validator with corrupted storage after an unclean shutdown; version skew between the querying client and one validator.

Related errors

AI-assisted analysis of linera-io/linera-protocol@6c226ddcb3 (2026-08-22). Data as JSON: /api/errors/469d9f0801a26ad4. Report an issue: GitHub.