Document Type | Technical Information
Category | Backup/Recovery
Document Number | TBATI025
Overview
This explains the meaning of the Down mode performed in Tibero and whether recovery is necessary depending on the type.
Method
Whether Recovery is Performed (at next boot) | Type of Down |
|---|---|
Not performed | Normal shutdown: normal, post_tx, immediate, switchover |
Performed | Abnormal shutdown: abort, abnormal |
NORMAL
This is the typical shutdown mode.
- After connecting to Tibero as the SYS user, it waits until all other sessions are disconnected before shutting down the server.
- After executing tbdown, no users can connect to the database anymore. However, users who were connected before tbdown was executed can continue to use the database until they disconnect themselves.
POST_TX
This mode waits until all transactions are finished before shutting down Tibero.
- Once tbdown POST_TX starts, no new connections to the database are allowed, and no new transactions can be started in existing sessions. However, currently running transactions can continue without restriction until they commit or rollback, at which point the database connection is automatically terminated.
- After tbdown starts, no separate notification such as message sending is given to clients connected to the database about the server shutdown. Utilities like tbSQL will recognize that Tibero has shut down when the next command is executed.
IMMEDIATE
This mode forcibly terminates all ongoing work, rolls back all ongoing transactions, and shuts down Tibero.
- Tibero shuts down after rolling back all interrupted transactions.
- If there were long-running transactions, rolling them all back may take some time.
- Clients are unaware of Tibero shutdown, similar to POST_TX mode.
SWITCHOVER
This mode synchronizes the Standby DB with the Primary DB and then shuts down the Primary DB like NORMAL mode.
- When shutting down in Switchover mode, it operates the same as NORMAL mode, but additionally sends all redo logs generated on the Primary to the Standby before shutdown.
- If one of the Standbys is shut down and then started in Failover mode, that DB becomes the new Primary.
ABORT
This is a forced shutdown mode that forcibly terminates the Tibero process. After connecting as the YS user, Tibero's MONP process sends the OS forced termination signal to all processes.
- After server shutdown, system resources such as shared memory and semaphores, as well as log files and data files, may remain. Also, recovery during the next Tibero startup may take a long time.
- All SQL currently being processed are aborted, and rollback is not performed.
CautionABORT is recommended to be used only in the following limited cases:
When normal shutdown is not possible due to an internal error in Tibero
When hardware problems require immediate shutdown of Tibero
When emergencies such as hacking occur requiring immediate shutdown of Tibero
ABNORMAL
This mode forcibly terminates the server process using OS forced termination signals regardless of the current Tibero server status and without connecting to the Tibero server.
- This mode is used in emergencies and requires recovery during the next Tibero startup.
- Because Tibero is forcibly terminated, system resources in use may not be released. After server shutdown, system resources such as shared memory and semaphores, as well as log files and data files, may remain. Also, recovery during the next Tibero startup may take a long time.