Disable SQL operations in training thread (#14947)

This is a temporary change until we figure out
why sqlite degrades the performance of training
threads significantly. To keep the number of
changes to a minimum, a connection is still
opened/closed when initializing/finalizing the
ML component.

1 files changed, 4 insertions, 0 deletions

diff --git a/ml/ml.cc b/ml/ml.cc
index 405c03928d..05dc6514f9 100644
--- a/ml/ml.cc
+++ b/ml/ml.cc
@@ -707,12 +707,16 @@ ml_dimension_train_model(ml_training_thread_t *training_thread, ml_dimension_t *
     {
         netdata_mutex_lock(&dim->mutex);
 
+        // temporarily disable sqlite operations because they interfere with
+        // training scheduling on busy parents
+        #if 0
         worker_is_busy(WORKER_TRAIN_LOAD_MODELS);
 
         int rc = ml_dimension_update_models(dim);
         if (rc) {
             error("Failed to update models for %s [%u, %u]", rrddim_id(dim->rd), dim->kmeans.after, dim->kmeans.before);
         }
+        #endif
 
         worker_is_busy(WORKER_TRAIN_UPDATE_MODELS);