CM installation problems (cf_installation.report not correct?)

Hi,

I am currently installing CM on a fresh system with EP SP9 (java only) on HPUX itanium 64.

The problem started when the sapinst was to restart the sapsystem after the deployment of the .sda via SDM. It didn't come down properly and therefore it was not able to start it. (haven't yet identified what caused this)

If manually stop it (by resorting to kill), I can get it up and the installation continues.

It complains that the cf_installation.report is not found, though cf_installation.report.bak is found. The bak file contains :

A Configuration upload sequence has started at: Thu Sep 08 16:17:25 CEST 2005

Number of upgrades to be processed: 69

META_ONLY_CA: OK

bc.util.prjconfig: OK

com.sap.netweaver.bc.uwl: OK

kmc.util.core: OK

bc.protocol.prjconfig: OK

bc.sf.prjconfig: OK

bc.sf.service.prjconfig: OK

kmc.people.appl.presence: OK

kmc.util.sor: OK

kmc.people.shared.cpr: OK

bc.rf.prjconfig: OK

bc.rf.repository.service.prjconfig: OK

bc.rf.manager.prjconfig: OK

I.e. that 13 of 69 have complete successfully.

By using SDM manually, I can see that it seems that all 69 of them are deployed to the server allready, but since the file only says 13 sapinst is on the seconds step (Deploy CM and collab component) showing the message "Portal configuration upgrade (13 of 69 upgrades done. 0 with warnings. Time elapse: 1700 seconds) Wait...

The log shows two messages:

Portal deployment Monitor warning. Nothing to monitor. No remaining EPT files found

Portal deployment Monitor warning. Nothing to monitor. No remaining PAR files found

The sapinst_dev.log contains little new information:

TRACE [iaxxcwalker.cpp:296]

CDomWalker::processStep()

Executing installation step EP_KMC_MAIN|ind|ind|ind|EP60|ind|0|EP60|ind|ind|ind|EP60|ind|0|EP_INSTALL_CM|ind|ind|ind|ind|ind|0|DEPLOY_CM_PAR_FILES|ind|ind|ind|ind|ind|0|runDeploymentMonitor

TRACE [iaxxejsexp.cpp:208]

EJS_Installer::writeTraceToLogBook()

Portal Deploy Monitor: Checking deployment of 0 EPT files and 0 PAR files under /usr/sap/ED2/JC00/j2ee.

Any ideas on where to attack this ???

I figure I might manually undeploy all the sda, remove the /tmp/sapinst_dir/NW/CMC dir(which make problems if I try to start sapinst again, which I have done quite a few times and start over

Dagfinn

BTW current script is :

                var fMgr = new FileMgt();
          var webas = WebAS.restore(context.get("WebASDump"));
          var j2eePath = "";
          var epts    = [];
          var pars    = [];
                
                for (var it = webas.getMyInstancesOnHost(); ! it.isDone(); it.next())
                {
                        var instance = it.get();
                        if ( ( instance.isJCI() || ( instance.isCI() && instance.hasJ2EE() ) ) )
                                j2eePath = instance.getDirInstance() + "/j2ee";
                }
                
                //  Switched to visitor mechanism to avoid fielnode invalidity problems
                
                var visitor = new MyVisitor( "ept" );
                gui.updateMsg( "ep.prt.executeStep", "Checking for undeployed EPTs in " + j2eePath + "/cluster/server0/apps/sap.com/irj/servlet_jsp/irj/root/WEB-INF/deployment/pcdContent" + "." );
    fMgr.getNode( j2eePath + "/cluster/server0/apps/sap.com/irj/servlet_jsp/irj/root/WEB-INF/deployment/pcdContent" ).takeVisitor(visitor);
    epts = visitor.getFileNames();

    visitor = new MyVisitor( "par" );
    gui.updateMsg( "ep.prt.executeStep", "Checking for undeployed PARs in " + j2eePath + "/cluster/server0/apps/sap.com/irj/servlet_jsp/irj/root/WEB-INF/deployment/pcd" + "." );
    fMgr.getNode( j2eePath + "/cluster/server0/apps/sap.com/irj/servlet_jsp/irj/root/WEB-INF/deployment/pcd" ).takeVisitor(visitor);
    pars = visitor.getFileNames();

                
                var mon  = new EpDeployMonitor( j2eePath, epts, pars );
                mon.start();

Message was edited by: Dagfinn Parnas