Due to the configuration of oracle restart, the database instance will automatically start to read-only mode after it goes down. Data synchronization needs to be started manually. Several failures are at night. Before finding the cause, add zabbix to monitor the adg synchronization status. When the synchronization stops, it will be issued. Early warning for timely processing.

Check the database log and found the following error:

ORA-04021: timeout occurred while waiting to lock object
LGWR (ospid: 9817): terminating the instance due to error 4021

Query mos, suspected oracle bug, oracle gave a repair method, increase the shadow parameter: ALTER SYSTEM set “_adg_parselock_timeout”=550 scope=both

So adjust the parameters, restart the database, observe for a period of time, the downtime does not appear again, hereby record.