Wrong hostname setting after Sun Cluster failover

[Former Member]

Hi Gurus,

our PI system has been setup to fail over in a sun cluster with a virtual hostname s280m (primary host s280 secondary host s281)

The basis team set up the system profiles to use the virtual hostname, and I did all the steps in SAP Note 1052984 "Process Integration 7.1 High Availability" (my PI is 7.11)

Now I believe to have substituted "s280m" in every spot where previously "s280" existed, but when I start the system on the DR box (s281), the java stack throws erros when starting. Both SCS01 and DVEBMGS00 work directories contain a file called dev_sldregs with the following error:

Mon Apr 04 11:55:22 2011 Parsing XML document.
Mon Apr 04 11:55:22 2011 Supplier Name: BCControlInstance
Mon Apr 04 11:55:22 2011 Supplier Version: 1.0
Mon Apr 04 11:55:22 2011 Supplier Vendor: 
Mon Apr 04 11:55:22 2011 CIM Model Version: 1.5.29
Mon Apr 04 11:55:22 2011 Using destination file '/usr/sap/XP1/SYS/global/slddest.cfg'.
Mon Apr 04 11:55:22 2011 Use binary key file '/usr/sap/XP1/SYS/global/slddest.cfg.key' for data decryption
Mon Apr 04 11:55:22 2011 Use encryted destination file '/usr/sap/XP1/SYS/global/slddest.cfg' as data source
Mon Apr 04 11:55:22 2011 HTTP trace: false
Mon Apr 04 11:55:22 2011 Data trace: false
Mon Apr 04 11:55:22 2011 Using destination file '/usr/sap/XP1/SYS/global/slddest.cfg'.
Mon Apr 04 11:55:22 2011 Use binary key file '/usr/sap/XP1/SYS/global/slddest.cfg.key' for data decryption
Mon Apr 04 11:55:22 2011 Use encryted destination file '/usr/sap/XP1/SYS/global/slddest.cfg' as data source
Mon Apr 04 11:55:22 2011 ******************************
Mon Apr 04 11:55:22 2011 *** Start SLD Registration ***
Mon Apr 04 11:55:22 2011 ******************************
Mon Apr 04 11:55:22 2011 HTTP open timeout     = 420 sec
Mon Apr 04 11:55:22 2011 HTTP send timeout     = 420 sec
Mon Apr 04 11:55:22 2011 HTTP response timeout = 420 sec
Mon Apr 04 11:55:22 2011 Used URL: http://s280:50000/sld/ds
Mon Apr 04 11:55:22 2011 HTTP open status: false - NI RC=0
Mon Apr 04 11:55:22 2011 Failed to open HTTP connection!
Mon Apr 04 11:55:22 2011 ****************************
Mon Apr 04 11:55:22 2011 *** End SLD Registration ***
Mon Apr 04 11:55:22 2011 ****************************

notice it is using the wrong hostname (s280 instead of s280m). Where did I forget to change the hostname? Any ideas?

thanks in advance,

Peter

[Former Member]

We finally resolved this. It turns out that the Unix SA had made a bad configuration in the Sun Cluster software (wrong username/password).

The error I found above still exists, but it is not affecting the proper function of the system, so I am ignoring it.

thanks anyway for your thoughts!

Peter

[Former Member]

by the way, the end result of these errors is that I cannot access the runtime workbench nor s280m:50000/dir to start the development tools.

the default traces are also full of errors accessing the gateway on host sun280.

thanks again!

[Former Member]

Please note that the PI system is transparent about the Failover system used.

When you configure the parameters against the mentioned note, this means that in case one of the nodes is down, the load will be sent to another system under the same Web Dispatcher/Load Balancer.

When using the Solaris failover solution, it covers the whole environment, including the web dispatcher, database and all nodes.

Therefore, please check the configuration as per the page below, which talks specifically about the Solaris failover solution for SAP usage:

http://wikis.sun.com/display/SunCluster/InstallingandConfiguringSunClusterHAfor+SAP