risingwavelabs/risingwave · error · anyhow::Error

reschedule failed

Error message

reschedule failed

What it means

The reschedule command submitted a fragment reschedule plan to the meta service but the reported success flag was false, so the command aborts with this error after printing the current plan revision.

Source

Thrown at src/ctl/src/cmd_impl/meta/reschedule.rs:132

            println!("\tChange: {:?}", reschedule.get_worker_actor_diff());
        }

        println!();
    }

    if !dry_run {
        println!("---------------------------");
        let (success, revision) = meta_client
            .reschedule(reschedules, revision, resolve_no_shuffle)
            .await?;

        if !success {
            println!(
                "Reschedule failed, please check the plan or the revision, current revision is {}",
                revision
            );

            return Err(anyhow!("reschedule failed"));
        }

        println!("Reschedule success, current revision is {}", revision);
    }

    Ok(())
}

// It will match formats like `1:[1:+1,2:-1,3:1];2:[1:1,2:1]`, indicating which workers' actors need to change in quantity for each fragment.
fn parse_plan(mut plan: String) -> Result<HashMap<u32, PbWorkerReschedule>> {
    let mut reschedules = HashMap::new();
    let regex = Regex::new(r"^(\d+):\[((?:\d+:[+-]?\d+,?)+)]$")?;
    plan.retain(|c| !c.is_whitespace());

    for fragment_reschedule_plan in plan.split(';') {
        if fragment_reschedule_plan.is_empty() {
            continue;
        }

View on GitHub (pinned to 6469eb736d)

Solutions

1. Re-run the reschedule with the current plan revision (fetch the latest revision first)
2. Adjust the plan to reference existing worker ids and valid fragment ids
3. Retry after the cluster stabilizes if a concurrent failover invalidated the plan

let latest = get_current_revision(&meta).await?;
if plan_revision != latest {
    return Err(anyhow!("stale plan revision {} != {}", plan_revision, latest));
}

When it happens

Trigger: Running `risectl meta reschedule` with a plan that conflicts with the cluster's current revision, or a plan the meta service rejects (e.g. moving actors for a worker that no longer exists).

Common situations: Cluster topology changed between composing the plan and applying it (node failure/scale-down); stale revision from a previous reschedule; invalid worker ids in the plan.

Related errors

• duplicate worker id {worker_id} in plan, prev {worker_id} ->
• Invalid time: {value} {unit} is out of range for a time of d
• Invalid jsonb encoding
• Invalid variant encoding
• should have meta client

AI-assisted analysis of risingwavelabs/risingwave@6469eb736d (2026-09-11). Data as JSON: /api/errors/2a0d4be7cf0afb71. Report an issue: GitHub.