DBA from Crisis to Confidence

Make sure the following patterns exist in the trace file: 数据挖掘实验室

- call stack

*** 2005-09-23 14:28:40.028
ksedmp: internal or fatal error
----- Call Stack Trace -----
calling              call     entry                argument values in hex
location             type     point                (? means dubious value)
-------------------- -------- -------------------- ----------------------------
ksedst()+27          call     ksedst1()            1 ? 1 ?
ksedmp()+557         call     ksedst()             1 ? 1 ? 5 ? F4166F ? F4E518 ?
....

- process state dump

===================================================
PROCESS STATE
-------------
Process global information:

Since it is 10g, the ASH should be discussed. It will have the past active waits.

...
 
数据挖掘研究院
 

d. Collecting log files in RAC environment:

When you have to log a Service Request for RAC Oracle environment, there are a few more log files you have to collect. Following is a list of files you may want to upload every time you log a Service Request.

• RDA output from all cluster nodes. An alternative is to always keep the recent RDA in proactive SR. Remember to renew RDA stuff each time there is a change in the environment – for instance, when a database or ASM persistent configuration parameter is changed.
• Database instances and ASM alert log files from all cluster nodes. The important part here is not to forget ASM logs – it well may be an ASM instance that is responsible for problems with the database proper.
• CRS log files from all cluster nodes. The collection procedure varies a bit depending on whether you are in 10gR1 or 10gR2 environment.
  • 10gR1: collect log files from $CRS_HOME/crs/log, $CRS_HOME/css/log, $CRS_HOME/evm/log, $CRS_HOME/crs/racg/dump.
  • 10gR2: collect log files from $CRS_HOME/log/`hostname` directory – you may want to ′tar′ the entire directory altogether rather than go through every subdirectory in there.