For DBA, the database is turned off and restarted to optimize, and the adjustment of the application is often encountered. If the user has logged as a database, you use the shutdown immediate or shutdown abort command to perform the shutdown database, the user will not be able to connect until the database restarts, the user will often complain: how to restart again. In fact, every time this start-up is made to be better for database. This article will specifically introduce the steps and many precautions that should be thought of when ShutDown / Startup operations, maybe these can help you. Note: In the first step, the SHUT Down database, do not turn off the SQL * NET in advance until you confirm that the database closes the database on the Server and then close the SQL * NET. Because once you close SQL * Net, users will lose and database sessions. The first step --- archive or delete old Trace Files and Logs When you start an instance of Oracle (instance), Oracle writes the diagnostic information about this instance to the specified trace and log file. Each background process adds a trace file that is stored in the specified directory (this path is specified by the background_dump_dest parameter in the init.ora or config.ora file). When the user encounters an internal error in the database, it also adds the user_dump_dest parameter in the init.ora or config.ora file when encountered an internal error in the database. The number and name form of the Trace file are determined by your operating system. For example, under UNIX, the file name is a digital plus drawing line plus the background process number, similar to 3_12345.trc, when you start the instance next time, a new file will be generated, because the process number is different, the file name may be 3_13245.TRC. Since each time you start and turn off the database, Oracle does not automatically delete the Trace file, so if you ignore this step, these useless files will occupy a lot of free space. So DBA should actively manage these files, or delete or archive in other specified directories, how to manage is mainly based on your own development environment. SQL * NET also generates a log file, generally called listener.log, in order to know his location, you can type in the command of the command lsnrctl status listener.log file size and the number of startups continues to grow, so it can be effectively managed Utilization space. Step 2 --- Remove the name Alert log file Alert log file is also used to record instance of diagnostic information. Its storage directory is also specified by the Background_Dump_Dest parameter. Under normal circumstances, the Alert log file is named: Alert_.log. The AlTer log file includes some database primary event information, such as the use of Startup, DHUTDOWN, Redo Log, the change of TableSpace, changes in the file, the state of internal error message, TableSpace-backup, etc.. The ALTER log file is important, and the size of the size of the file is also very amazing. If you don't check him, you want to see it after a while, because it is already very big, it is also costly. In order to leave these useful information, it is not a large space to be wasted. You can change the Alter Log file name at any time, even when Database Open. When Oracle can't find the file specified by Background_Dump_Desc, he will add a new Alter Log file.
