cancel
Showing results for 
Search instead for 
Did you mean: 

Central Instance failed over automatically

Former Member
0 Kudos

Hi experts,

We have an ECC 6 with EHP4 running on AIX 6.1 with VCS cluster. Last week we had an issue where Central Instance Failed over to secondary node automatically. I could not see any thing evident in Work logs or the VCS (engine_A.log) which caused/initiated this failover.

I am looking for some pointers to logs which can show some information related to this. Below are few things I have checked:

1) VCS - Engine_A.log -  This shows that VCS is looking for all the Process ID's. Then it couldnt find some process and initiated a failover. It also says bringing resource group offline (not initiated by VCS).

2) dev_disp : this shows SoftCancel request for T69 U70 M0 received from DISPATCHER .

Apart from disp log I couldnot see any other information from work directory.

Additionally Oracle database is also running on same node which didnt had any problem durign this.

Please let me know if there is any thing else I can look for..

Accepted Solutions (0)

Answers (1)

Answers (1)

former_member189725
Active Contributor
0 Kudos

Probably you message server was unreachable , check the dev_ms trace file in the ASCS working directory. The dispatcher could not communicate with the message server , resulting in shutdown of the workprocesses. Check if the network had any issues probably something  intermittent during the failover time.

Regards

Ratnajit

Former Member
0 Kudos

Hi Ratnajit,

I can see a disconnection message in dev_ms trace but that is after the central instance went down. So I ruled out that possibility.

csaba_goetz
Contributor
0 Kudos

Hello Basis Team,

Check the OS syslog and sapstart.old if it exists. sapstartsrv.log (after restart it is renamed as sapstartsrv.old) could contain some useful information. Please attach the VCS - Engine_A.log (in which you could see that failover was initiatiated), sapstart.old and AIX syslog as well.

Best regards,

Adam

Former Member
0 Kudos

Hi Adam,

I have attached engineA.log and sapstartsrv.old  files.

Former Member
0 Kudos

Attached

ken_halvorsen2
Active Participant
0 Kudos

Hi Basis Team

Did you ever find out what caused this situation? I'm having the same problem just start up with CRM2007 with SQL Server 2005 on an MS Cluster.

SAP is fine to start up and run on 1 node but not the other. The database can start up and run on either node.

In the dev_disp log it has:

DpSigInt: caught signal 2

DpHalt: shutdown server >xxx < (normal)

DpModState: Chneg server from Active to Shutdown

Stop work processes

Ken