RE: RAC on windows 2003 - 32-bit - nightmare

Shameless self promotion:
   
  I was in you shoes about one year ago (except not on Rac).  I wrote an 
article about issues on Metalink and published it to Metalink. 427993.1. Look 
for suggestions on /3G switch.
   
  I would start from the ground up to make sure the basic configurations of 
both nodes are ok and the scp/rcp is set up correctly, interconnects, 
listeners, voting disks etc.
   
  I would do some basic transfers from one box to the other and see how those 
times look
   
  There is also a cluster utility (cvu,  on Linux at least to check to see if 
all the setups for Rac are done correctly.  I would assume Windows would have a 
utility as wel
   
  Is the second server have good response when it is running alone as well?
   
  Check your sqlnet.ora settings on each server.  Is there a timeout set there?
   
   
  Kathy

"Crisler, Jon" <Jon.Crisler@xxxxxxx> wrote:
        v\:* {behavior:url(#default#VML);}  o\:* {behavior:url(#default#VML);}  
w\:* {behavior:url(#default#VML);}  .shape {behavior:url(#default#VML);}        
        1)       Make sure that the network interconnect is negotiating at 
1gigabit full duplex.  Don?t accept the answer that ?the hardware supports it? 
but VERIFY that it is negotiating at that speed.
  2)       Ora-4031 ? not enough memory in shared pool.
  3)       Global cache issues:  again, network issues can affect this.  Make 
sure all sequences have cache values, as this can hurt performance and is often 
overlooked. 
   
      
---------------------------------
  
  From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On 
Behalf Of sam roberts
Sent: Saturday, September 29, 2007 11:10 PM
To: oracle-l
Subject: Fwd: RAC on windows 2003 - 32-bit - nightmare

   
   
    I have just inherited a RAC 2 node database: Using DELL Poweredge 2950 
Servers with 16GB RAM and DELL power vault MD3000 SAS Storage:

    I have previously absolutely no experience on Windows, as Im UNIX and LINUX 
and as I mentioned I inherited this built.

     

    The problem is that when both Nodes are up and running there are many 
problems:

    1)running a 5 minute export takes 5 hours and times out - session_wait says 
busy_buffer

    2)ora600 and ora-04031 errors on client and eventually cant log in

    3)tns12170 errors intermittently when logging in

    4) lots off warnings of global_cacahe_wait excessive:

     

    When we switch one node of the performance and everything runs normally. I 
have been told this is the 3rd build of this system and it happened everytime.

    I am suspecting the network right config right now.

     

    Please any suggestions from you guys would be highly appreciative.




       
---------------------------------
Moody friends. Drama queens. Your life? Nope! - their life, your story.
 Play Sims Stories at Yahoo! Games.

Other related posts: