clockworklabs/SpacetimeDB · critical · Error

transaction retry failed again

Error message

transaction retry failed again

What it means

Procedures run inside anonymous transactions via runWithTx: the body executes, then sys.procedure_commit_mut_tx() commits. If the first commit throws, the runtime logs 'committing anonymous transaction failed', replays the whole transaction once, and commits again; a second commit failure throws 'transaction retry failed again' with the underlying host error attached as `cause`.

Source

Thrown at crates/bindings-typescript/src/server/runtime.ts:367

      sys.procedure_abort_mut_tx();
      throw e;
    }
  };

  let res = run();
  try {
    sys.procedure_commit_mut_tx();
    return res;
  } catch {
    // ignore the commit error
  }
  console.warn('committing anonymous transaction failed');
  res = run();
  try {
    sys.procedure_commit_mut_tx();
    return res;
  } catch (e) {
    throw new Error('transaction retry failed again', { cause: e });
  }
}

type FlatSubmoduleDispatch = {
  reducerFns: Reducers;
  reducerDefs: RawReducerDefV10[];
  procedureFns: Procedures;
  procedureDefs: RawProcedureDefV10[];
  anonViewFns: AnonViews;
  viewFns: Views;
  tables: Array<{ accessorName: string; tableDef: RawTableDefV10 }>;
  schemaTables: Record<string, UntypedTableDef>;
  typespace: Typespace;
  dbView_: DbView<any> | undefined;
  queryBuilder_: QueryBuilder<any> | undefined;
  /** e.g. "alias." for a submodule with namespace alias "alias" */
  namePrefix: string;
  subDispatches: SubmoduleDispatchInfo[];

View on GitHub (pinned to 524b4487d9)

Solutions

1. Inspect error.cause: it carries the real sys error that made both commits fail
2. If the cause is deterministic (constraint or budget violation), fix the procedure body; a blind retry cannot succeed
3. If the cause looks transient/host-level, check spacetimedb host logs and resource limits for the instance
4. If the data should legitimately commit, report an issue with the cause attached and the module hash

try {
  await runProcedure();
} catch (e) {
  const cause = (e as Error)?.cause ?? e;
  if (isTransientHostError(cause)) {
    scheduleRetryWithBackoff(); // procedure must be idempotent
  } else {
    logAndAlert('procedure commit failed', { cause });
  }
}

When it happens

Trigger: Any host-side commit error occurring twice in a row: a datastore error at commit time, resource/budget exhaustion, or a deterministic failure caused by the procedure's own writes (a deterministically failing write fails identically on replay).

Common situations: Module under load hitting host limits; writes that violate host-level constraints surfacing only at commit; transient host storage issues where the single built-in retry is not enough.

Related errors

• unknown procedureId ${id}
• wrong number of elements
• too many elements
• cannot serialize refs without a typespace
• cannot deserialize refs without a typespace

AI-assisted analysis of clockworklabs/SpacetimeDB@524b4487d9 (2026-08-16). Data as JSON: /api/errors/4df230e9e01e5cd7. Report an issue: GitHub.