From 0240d91743856fe945c4fcab5e88a7c6855f707c Mon Sep 17 00:00:00 2001
From: Ferran Llamas <ferran@nuclia.com>
Date: Thu, 20 Jun 2024 17:55:00 +0200
Subject: [PATCH] Only configure partitions on ingest statefulset command
 (#2266)

---
 nucliadb/src/nucliadb/ingest/app.py               |  3 +--
 nucliadb/src/nucliadb/search/search/paragraphs.py | 13 +++++++++++--
 2 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/nucliadb/src/nucliadb/ingest/app.py b/nucliadb/src/nucliadb/ingest/app.py
index 4d9d640dda..03daabb789 100644
--- a/nucliadb/src/nucliadb/ingest/app.py
+++ b/nucliadb/src/nucliadb/ingest/app.py
@@ -175,8 +175,6 @@ async def main_subscriber_workers():  # pragma: no cover
 def setup_configuration():  # pragma: no cover
     setup_logging()
 
-    assign_partitions(settings)
-
     errors.setup_error_handling(importlib.metadata.distribution("nucliadb").version)
 
     if asyncio._get_running_loop() is not None:
@@ -190,6 +188,7 @@ def run_consumer() -> None:  # pragma: no cover
         - pull worker
     """
     setup_configuration()
+    assign_partitions(settings)
     asyncio.run(main_consumer())
 
 
diff --git a/nucliadb/src/nucliadb/search/search/paragraphs.py b/nucliadb/src/nucliadb/search/search/paragraphs.py
index ae1dc2459d..da4f3048a9 100644
--- a/nucliadb/src/nucliadb/search/search/paragraphs.py
+++ b/nucliadb/src/nucliadb/search/search/paragraphs.py
@@ -125,7 +125,13 @@ async def get_paragraph_from_full_text(
     """
     extracted_text = await get_field_extracted_text(field, cache=extracted_text_cache)
     if extracted_text is None:
-        logger.warning(f"{field} extracted_text does not exist on DB yet")
+        logger.warning(
+            "Extracted_text for field does not exist on DB. This should not happen.",
+            extra={
+                "field_id": field.resource_unique_id,
+                "kbid": field.kbid,
+            },
+        )
         return ""
 
     if split not in (None, ""):
@@ -154,7 +160,10 @@ async def get_paragraph_text(
     if orm_resource is None:
         orm_resource = await get_resource_from_cache(kbid, rid)
         if orm_resource is None:
-            logger.error(f"{kbid}/{rid}:{field} does not exist on DB")
+            logger.warning(
+                "Resource does not exist on DB. This should not happen.",
+                extra={"resource_id": rid, "kbid": kbid, "field": field},
+            )
             return ""
 
     _, field_type, field = field.split("/")