Error getting confirmation reply during attach to cluster

Question

Hi Experts,
Our portal system has an issue.
Symptoms are as follows
Portal response time increases and availability goes down (some times on peak load)
CPU utilization becomes 100% on application servers.
We try to restart the application instances 1 after another.
And we get following errors in std_serverX.out files.
while restarting the server 13 we are getting the Fatal Error :
Loading: ClusterManager ... Framework -> CriticalShutdown Error getting confirmation reply during attach to cluster.
Jan 11, 2010 4:20:11 PM com.sap.engine.core.Framework [ThreadThread-1,5,main] Fatal: Critical shutdown was invoked. Reason is: E
rror getting confirmation reply during attach to cluster.
Then we have to stop all application servers.
Then we restart the SCS instance.
After SCS restart,Portal starts normally without above error.
This happens only when there is performance issue.
if we try to restart the portal application servers 1 after another on week ends then it does not throw above error.
Kindly suggest how can I avoid above error while performing restart 1 after another.
It would allows us to maintain availability of portal while performing restart 1 after another.
Thanks and Regards,
Girish

Answer 1

Hi,
There could be two reasons for this.
01. The network connectivity between central instance to application servers. Check the utilization and take corrective action.
02. Due to excess memory, your swap is taking more time in cleaning the heap size. Hence you can add more nodes or reduce excess memory.
As an work around, you can check the value of the following parameters and increase it to avoid the restart.
ms.notification.timeout and ms.confirmation.timeout
These parameters have to be set in configtool.
Pls let me know if this resolves your problem.
Regarsd,
Iyyappan M R