Oracle Background Process Feature description

I. Smon: System monitoring Process
1. Mainly submitted data but has not yet been written to the data file to power down, restart the time, the Smon will be restored
A. Roll forward, the redo log file has been written but no data has been written to the data file (using SCN numbers to identify the commit record)
B. Open the database
C. Rolling back a transaction that is not committed
2. Disk space maintenance can also be done
A. Reclaim or merge idle areas connected to data files
B. Releasing a temporary segment
ii. Pmon: Process monitoring Process
1. When a process crashes, Pmon will be responsible for the following cleanup work
A. Rolling back a user's current transaction
B. Releasing all table-level and row-level locks added by the user
C. Releasing all other resources of the user, etc.
Three, CKPT: check (check) point and inspection point process
1. Improve system efficiency and database consistency, introduce a checkpoint event, CKPT is when DBWR writes data from the cache to the data file
It is not needed in theory.
Full recovery can be ensured through the redo log and SCN, which is introduced for efficiency, as all changes to the check point have been written to the number
According to the document, at the time of recovery
The Redo log before the check point is not needed, so that the instance recovery is accelerated, CKPT will officer to all relevant data files in the file header, but also officer the checkpoint number, redo log sequence number, archive log name and the lowest, maximum SCN number is written to the control file, because CKPT Generates a large number of IO operations, so it is not possible to generate checkpoints frequently, preferably at least 20 minutes
iv. ARCn: Archive log Process
1. When a group of redo log files are full, another group is written, and when the last group is full, the first group is written, called Log switching
2. If the archive is not open, then it cannot be restored after overwriting.
3. If the archive is open, the arch process is used to write to the archive log file, and Oracle ensures that the archive operation is completed in a set of redo logs
The group log will not be reused before
v. DBWN: Database Write Process
Before 1.11g can only have 0~9,10 process, 11g can have 36, 0~9 and A~z, through the db_writer_processes to set the number
If no number is stated when starting the
Then the number of CPUs to determine the parameters of the
2. DBWR writes data from the buffer zone to the data file when the following conditions occur
A. When the number of dirty buffers exceeds the set limit
B. The time interval has been set to
C. There is a process that requires a high-speed buffer zone, but there is no free buffer space
D. When a check (checkpoint) occurs
E. A table is deleted or truncated (Truncate,drop)
F. A table space is set to read-only
G. When a tablespace is backed up online
H. Table space is set to off-machine
vi. LGWR: Redo Log Write Process
1. Write the record order of the redo log buffers to the redo log file, noting that it is written sequentially, so the IO is relatively small
2. How the Commit statement works, using the quick submit technique
A. The server process writes the submitted record along with the resulting SCN number to the redo log buffer
B.lgwr writes the records in the redo log buffer to the redo log file, which ensures that the
C.oracle notifies the user (process) that the submission has been completed
E. The server process modifies the state of the relevant data in the database cache and frees the resource and open the lock (this may not be the data written
To the data file, they are marked as dirty data, and then written to the data file by DBWR
3. LGWR will write to redo log file in the following situation
A. Transaction submission
B. Record in redo log buffer exceeds 1/3 of buffer capacity
C.DBWR data blocks in the database cache are written to the data file
D. Every 3 seconds

Oracle Background Process Feature description