Fix: Use explicit UID 105 for pgdata ownership in sync job

The copy_datadir.sh script uses 'postgres:postgres' which resolves to the
CI runner's postgres UID (70 on some systems). However, the HAF container
expects postgres to be UID 105. This mismatch causes PostgreSQL to fail
because it cannot read the pgdata directory.

Fix by explicitly chowning haf_db_store to 105:109 after copy_datadir.sh
runs, ensuring the HAF container's postgres user can access the data.