Oracle control files and some concepts related to log files

Redo01.log, redo02.log, and redo03.log are online files, which are first stored in logs. After full storage, they are stored in redo02.log and then in redo03.log. After the storage is complete, they are stored in redo01.log cyclically (covering: the structure of www.2cto.com is as follows:

2. The control file's importance to Oracle is a very important file for Oracle databases, just like the windows registry. If the control file is accidentally damaged, the Oracle system may fail to start normally. Therefore, it is necessary to ensure the security of control files during routine maintenance and management of Oracle databases. In practice, the Oracle administrator can back up the control file to improve the security of the control file. However, when the control file is corrupted and recovered through the backup file, the Oracle database will be shut down for a period of time. Therefore, it is best to use multiplexing to ensure the security of control files. When multiplexing is used, when a control file is damaged, the system will automatically enable another normal control file to start the Oracle database, so there will be no downtime.
The multiplexing feature of the control file www.2cto.com: the principle of multiplexing is actually very simple, that is, the control file is stored in multiple disk partitions or multiple hard disks on the Oracle Database Server. When the control file needs to be updated, the Oracle database system automatically updates multiple control files at the same time. When one of the control files is corrupted, the system automatically enables another control file. Only when the Oracle database administrator is lucky enough, that is, all the control files are damaged. In this case, the Oracle database cannot be started normally and needs to be started manually, causing some loss. However, as long as the control file is reused on multiple hard disks, the damage to the control file is usually very small. Therefore, the use of Multiplexing Control files can greatly improve the security of control files. Most importantly, there will be no downtime in the Process of controlling file conversion. Therefore, Oracle databases can avoid the risk of spof by storing control files on different hard disks. When multiple control files are enabled using multiplexing technology, Oracle databases update control files at the same time. Although sometimes this may allow the Oracle database to spend more time modifying the control file, the investment in this time is worthwhile relative to the risk. In addition, as long as the performance of the hard disk is good, the impact of this time is very small, or even negligible. However, when multiplexing is adopted, it is best not to prevent control files from being placed on servers on the network. Sometimes, if the system encounters poor network performance or even network disconnection when updating the control file, it takes a long time to update the control file. Www.2cto.com 3. Difference Between redo log and archive log: redo log file is written by the LGWR process from the redo log buffer in the Oracle instance and is recycled. That is to say, a redo log file (group) is written to the next one only after it is fully written. Archive log: When the database is running in archive mode, when a redo log file (group) is full, the ARCn process backs up the redo log content to the archive log file, then the redo log file (group) can be used for the next time. Whether or not the database is in archive mode, redo logs must be written. Only when the database is in archive mode can the redo log be backed up to form an archive log. Generally, archive logs are combined with full backup for recovery after database problems occur. Www.2cto.com redo logs are recycled. For example, there are three redo log Groups A, B, and C. Then, when A is full, the system calls the ARCn process to back up A as an archive log, and B has started to use it. Assume that you only have two groups, A and B. In some cases, A is backing up and cannot be used until it is finished. B is full. At this time, the database will be suspended. Therefore, in general, it is best to redo the log in three groups or a little more, and the size should be appropriate. In fact, when a redo log group is full, archive logs are written. If ABC is full and archive again, it must be suspended. This is not the case for Oracle. Both the archive log and redo log are physical files, but they are stored in different directories, and the file name of the redo log remains the same. The file name of the archived log is generated by the system during Backup. After the redo log is backed up as an archive log, the system clears the redo log content, but the file still exists and is ready for use next time.