Home » Server Options » RAC & Failsafe » RAC Performance Problem - Hang (HP-UX Oracle 10.2.0.1)
RAC Performance Problem - Hang [message #405855] Sat, 30 May 2009 06:46
katykpr
Messages: 1
Registered: May 2009
Junior Member
Dear Sir/Madam

The following is a brief specification of our system :

1-TPS : 1000-1700
2-DB Size : about 1TB
3-Concurrent connections : 1500-2000 (Active+ Inactive)
4-2 Nodes RAC _ 10.2.0.1.
5-OS : HP/UX

Performance of the system is generally accepted unless something undesirable happen in system like : 'Running resource consuming reports-FTS, Network latency ,...' , though clearly the root cause need to be addressed ,
Yet ,
The main problem is even after removing the root cause (killing culprit SQL ,...) the whole system (specially RAC interconnect ) face severe problems 'Many "gc_%" events occure' as such that the problem could not be solved until one of the nodes restart .
Because the system is mission critical and no downtime is acceptable , managing difficulty with such manner is neither accepted nor feasible , hence the question is why even after killing the problematic session and sometimes all sessions ,the system is completely out of control until node restart.
Why system could not handle the situation gracefully , surprisingly after such situation , the system work fine with one node !

Thanks in advance
KPR


Previous Topic: Dataguard vs RAC
Next Topic: Nodes Get Rebooted
Goto Forum:
  


Current Time: Thu Mar 28 12:06:55 CDT 2024