Yesterday, I was working with Oracle’s data pump export tool, which is something I don’t normally do. I am not really sure why Oracle is so difficult to work with in regards to backing up, migration of data and recovery of data. I mean it is just flat out confusing when you don’t do it on a regular bases.
So I ran into all sorts of weird errors. The first one was:
ORA-23603: STREAMS enqueue aborted due to low SGA – after some research I found that the my SGA area was too small, which in my opinion “Who Cares” you are simply dumping data to a flat file why in the world does that need to be complicated, but….
The quick google search turned up to increase the streams_pool_size parameter and increase the SGA memory target area. The first change I needed to make was to increase the sysctl parameter to as this parameter defines the maximum size in bytes of a single shared memory segment that a Linux process can allocate in its virtual address space. So I modified the sysctl.conf file and increased my kernel.shmmax paramter and set the change active with the sysctl -p call. You need to shutdown/start the database for this change to be made. I also modified the init.ora file to increase the streams_pool_size parameter.
It took a couple of tries to get everything to run correctly, however, the last run I did presented me with the UDE-00008 error. What the heck, everything looks great, so again another google search reveals a novel idea: check the logs:
Master table “SYSTEM”.”SYS_EXPORT_FULL_02? successfully loaded/unloaded
******************************************************************************
Dump file set for SYSTEM.SYS_EXPORT_FULL_02 is:
/oracle/datadump/fs92demo.dmp
Job “SYSTEM”.”SYS_EXPORT_FULL_02? successfully completed at Wed Jun 17 11:13:53 2015 elapsed 0 00:41:17
Well, what do we have here. Everything completed successfully. Apparently there is a glitch in Oracle’s data pump that sometimes appears randomly however, the job has ran successfully. Simply ignore the error and move on! If you want to read up on the bug apparently the bug number on Oracle is: 5969934. I honestly, didn’t take the time to read it.