fix: disable checkpointing

Persisting pipeline processors to disk currently uses a lot of disk space, especially when checkpointing frequency is high; for example when `max_num_records_before_persist` is set to a low value.

dozer-core/src/checkpoint/mod.rs

@@ -301,13 +301,18 @@ async fn read_record_store_slices(
     let record_store = ProcessorRecordStoreDeserializer::new(record_store)?;
     let record_store_prefix = record_store_prefix(factory_prefix);
 
-    let mut last_checkpoint: Option<Checkpoint> = None;
+    let last_checkpoint: Option<Checkpoint> = None;
     let mut continuation_token = None;
     loop {
         let objects = storage
             .list_objects(record_store_prefix.to_string(), continuation_token)
             .await?;
 
+        if let Some(_) = objects.objects.last() {
+            return Err(ExecutionError::CannotRestart);
+        }
+
+        #[cfg(FIXME_CHECKPOINTING)]
         if let Some(object) = objects.objects.last() {
             let object_name = AsRef::<Utf8Path>::as_ref(&object.key)
                 .strip_prefix(&record_store_prefix)

dozer-core/src/errors.rs

@@ -50,6 +50,8 @@ pub enum ExecutionError {
         source_name: NodeHandle,
         table_name: String,
     },
+    #[error("Dozer cannot restart. You have to clean data from previous runs by running `dozer clean`")]
+    CannotRestart,
     #[error("Failed to create checkpoint: {0}")]
     FailedToCreateCheckpoint(BoxedError),
 }

dozer-core/src/executor/processor_node.rs

@@ -115,6 +115,7 @@ impl ReceiverLoop for ProcessorNode {
             self.error_manager.report(e);
         }
 
+        #[cfg(FIXME_CHECKPOINTING)]
         if let Some(checkpoint_writer) = &epoch.common_info.checkpoint_writer {
             let object = checkpoint_writer.create_processor_object(&self.node_handle)?;
             self.processor