<fix>[ai]: add clusterId column to ModelServiceInstanceVO for node restart recovery

Add flyway migration to persist clusterId on ModelServiceInstanceVO so that
inference service restart can locate the K8s cluster even after Pod eviction.
Includes backfill of existing data from PodVO.

Resolves: ZSTAC-81566

Change-Id: I3136b558d35bf723b49203e4cdcf351c0cd33a37

Author: AlanJager

conf/db/upgrade/V5.5.6__schema.sql

@@ -190,3 +190,12 @@ WHERE p.type NOT IN (
     'GPU_Co_Processor',
     'GPU_Communication_Controller'
 );
+
+-- ZSTAC-81566: Persist clusterId on ModelServiceInstanceVO to survive node restart
+CALL ADD_COLUMN('ModelServiceInstanceVO', 'clusterId', 'BIGINT', 1, NULL);
+
+-- Backfill existing data from PodVO
+UPDATE ModelServiceInstanceVO msi
+INNER JOIN PodVO p ON msi.vmInstanceUuid = p.uuid
+SET msi.clusterId = p.clusterId
+WHERE msi.clusterId IS NULL AND p.clusterId IS NOT NULL;