oracle hang分析（Oracle挂分析）.doc

oracle hang分析（Oracle挂分析） Oracle Hang analysis Personal classification: Architecture chapter 1. database hang several possibilities Oracle deadlocks or system loads are very high, such as CPU usage or some other locks. Waiting high can cause system hang to live, such as a large number of DX locks. Generally speaking, the system hang we refer to means that the application has no response, the ordinary sqlplus can hardly operate, and so on. 2. how do you perform hang analysis? Which level does hang analysis have? How do I select level? Hanganalyze has the following level: 10, Dump, all, processes (IGN, state) 5, Level 4 + Dump, all processes, involved, in, wait, chains (NLEAF, state) 4 Level, 3 + Dump, leaf, nodes (blockers), in, wait, chains (LEAF, LEAF_NW, IGN_DMP, state) 3, Level 2 + Dump, only processes, thought, to, be, in, a, hang, (IN_HANG, state) Only, HANGANALYZE, output, no, process, dump, at, all, How do I select level? In general, it is not recommended to use more than 3 levels of Hang analysis, because it may produce very large trace, and may have an impact on the systems IO. From the start of Oracle hanganalyze, 9i provides support for rac. There are 2 ways: ()) ALTER, SESSION, SET, EVENTS,immediate, trace, name,, HANGANALYZE, level, level; 2) using the oradebug command ORADEBUG setmypid ORADEBUG setinst all ORADEBUG, -g, def, hanganalyze, level - - for RAC usage Oradebug setmypid Oradebug hanganalyze 3 - non RAC environment Usually, when doing hang analysis, Oracle suggests doing a systemstate dump at the same time Oradebug SYSTEMSTATE dump Level 2 Level 2, which contains all the information about session. Sqlplus -prelim / as SYSDBA ---10g can be logged in this way Oradebug setospid process ID Oradebug unlimit Oradebug dump systemstate 10 Added: sometimes we may need to have a trace AIX environment for a process, and we can use the DBX command Following example: DBX, -a, PID (where PID = any, Oracle, shadow, process) -- see through the PS -ef|grep XXX D