warpdotdev/warp · error

Failed to persist forked conversation at block: {e:?}.

Error message

Failed to persist forked conversation at block: {e:?}.

What it means

After building the forked conversation's tasks and data, the model persists them by sending `ModelEvent::UpdateMultiAgentConversation` (carrying the fresh forked_conversation_id, updated tasks, conversation data) over the `model_event_sender` channel to the SQLite layer. If `send` fails — the channel is closed or its receiver dropped — persistence aborts and this error is returned, so the fork does not survive as durable state.

Source

Thrown at app/src/ai/blocklist/history_model.rs:1928

            parent_agent_id: None,
            agent_name: None,
            orchestration_harness_type: None,
            parent_conversation_id: None,
            is_remote_child: false,
            root_task_is_optimistic: None,
            run_id: None,
            autoexecute_override: Some(conversation.autoexecute_override().into()),
            last_event_sequence: None,
            pinned: false,
        };

        let forked_conversation_id = AIConversationId::new();
        if let Err(e) = sqlite_sender.send(ModelEvent::UpdateMultiAgentConversation {
            conversation_id: forked_conversation_id.to_string(),
            updated_tasks: updated_tasks_with_new_ids.clone(),
            conversation_data: conversation_data.clone(),
        }) {
            return Err(anyhow!(
                "Failed to persist forked conversation at block: {e:?}."
            ));
        }

        let forked_conversation = self.insert_forked_conversation_from_tasks(
            forked_conversation_id,
            updated_tasks_with_new_ids,
            conversation_data,
        )?;

        Ok(forked_conversation)
    }

    pub fn apply_client_actions(
        &mut self,
        response_stream_id: &ResponseStreamId,
        client_actions: Vec<warp_multi_agent_api::ClientAction>,
        conversation_id: AIConversationId,

View on GitHub (pinned to e72fd7aacb)

Solutions

1. Retry the fork after restarting the app — the source conversation is untouched and only the fork is lost
2. Check logs for an earlier panic or shutdown of the persistence actor that closed the channel (that is the root cause, not the fork itself)
3. Disable fork/mutation actions during shutdown windows so sends cannot race teardown
4. If persistent, inspect how model_event_sender is wired in GlobalResourceHandlesProvider

// Skip mutation actions when the persistence channel is already closed.
if sqlite_sender.is_closed() {
    return Err(anyhow!("Persistence channel closed; retry after restart"));
}

if let Err(e) = sqlite_sender.send(ModelEvent::UpdateMultiAgentConversation { .. }) {
    log::error!("Failed to persist forked conversation: {e:?}");
    // safe to retry after restart: source conversation is unchanged; regenerate the fork id
    return Err(anyhow!("Failed to persist forked conversation at block: {e:?}"));
}

When it happens

Trigger: `sqlite_sender.send(ModelEvent::UpdateMultiAgentConversation { .. })` returns Err because the channel's receiver is gone: the DB/persistence actor shut down (quit race), panicked earlier, or the test harness never ran a receiver. Note the freshly generated `AIConversationId::new()` means a retry creates a different id.

Common situations: User forks a conversation exactly as the app quits or the persistence worker restarts; SQLite actor panicked on an earlier malformed event; tests constructing history_model without the model-event loop.

Related errors

• No sqlite sender available.
• Agent driver dropped while sending /exit
• Truncated tasks for forked conversation at block are empty f

AI-assisted analysis of warpdotdev/warp@e72fd7aacb (2026-08-16). Data as JSON: /api/errors/c7cad1b6a1e0aa66. Report an issue: GitHub.