Shulou(Shulou.com)06/01 Report--

Overview:

Verify the role of data guardian in real-time master and standby

Compare oracle dg data protection modes.

First of all, the real-time master and standby database status and archive status are modified by the data daemon.

What is the impact of standby library shutdown on the main library?

What is the problem when the daemon dies?

1. The daemon is intact and the standby library is closed.

There are two situations:

1. Abnormal shutdown of standby database

The daemon attempts to start the standby library:

1. Instance status error detected

two。 The daemon switches state from open to startup

3. Start recovery of instance (restart, instance recovery) recover to ok

4. The daemon switches to the unify ep unified node state.

5. The daemon then switches to the startup state, and finally open.

two。 Shut down via monitor

The daemon does not attempt to start the standby library

two。 Daemon modifies archive status

1. Close the standby library daemon and take care of the library (direct kill process)

The archive of the slave database will be changed to invalid, and the main database will be normal:

1. The master database guards the state from open to mon confirm, and the instance enters the suspended state.

two。 After the monitor confirms the slave exception, the daemon enters the failover state

3. The daemon goes from failover to open, and the instance hangs to open.

The main library can operate normally:

Archive the repository as invaild:

two。 Turn off the main library daemon, take care of the library (kill master library daemon, shut down the standby library through monitor)

The archive of slave database will not be changed to invalid, and the commit operation of master database will suspend the database:

Start the standby daemon:

1. The daemon goes from startup to unify ep state, and the instance starts to mount

two。 The daemon goes from unify ep to startup, and the instance starts to open

3. The daemon goes from startup to open.

Notify the main database daemon for data recovery

1. The main database daemon goes from open to recovery (archives during slave shutdown are sent to standby again)

two。 The main library daemon goes from recovery to open

Exception: simulated daemon exception

Kill the main database daemon manually, and shut down the slave instance under monitor:

Under monitor: kill instance dm1.dmstd

1. Slave instance is closed with a status of error

two。 The standby daemon enters the shutdown state.

At this point, the main library enters the suspend state and the database is unavailable.

Data can be queried and modified in suspend, but when the database returns to normal, all operations during suspend are rolled back.

Summary:

Data guardianship:

1. Monitor and automatically pull up the database (if the guardian fails monitor, the latest information of the database will not be found, the kill database process, the database will be started automatically)

two。 Modify the archiving status (if the slave is closed through monitor, the daemon will not start the database automatically.) the caretaker of the master database will change the archiving status to invalidation, so that the normal operation of the master database will not be affected. If the daemon of the primary library hangs, and then the standby database dies, triggering the operation of writing the redo log will cause the database to hang. When the daemon is restarted, the database automatically returns to normal and rolls back the operation during suspend.

Main library guardian

Prepare storehouse guardian

Main database instance

Prepare a database instance

Archiving status

Normal

Normal

Open

Automatic restart

Valid

Normal

Shutdown

Open

Close

Invalid

Kill

Shutdown

Suspend

Close

Valid

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.