Configuring Parallel Concurrent Processing with ssh

Question

Happy New Year,
Metalink Note 185489.1 describes setting up PCP so that the applmgr account on each node can rsh into the other apps tier nodes, but has anyone been successful in setting it up with ssh? I see that the last time this document was updated was in 2004, quite some time ago.
Sincerely,
Ted Persky
Sun Microsystems, Inc.

Answer 1

I have configured PCP in 11.5.10.2 having two application server with OS RHEL 4.0 as follows:
a. On node app1 check whether the apps_sid listener is up and if it is down bring it up.
b. Log in to the applications by system administrator responsibility and do the following steps:
1. Check whether both the nodes are registered by navigating to Install > Nodes and verify.
2. Navigate to Concurrent > Manager > Define and query for Internal Manager.
3. In the secondary node box enter the name of <node2> and save the changes.
4. Do the same for some other concurrent managers except Standard.
5. For some concurrent managers do the reverse i.e. enter aap2 as the primary node and app1 as secondary node.
6. Query whether service managers of the name Service Manager (<node2>) and Internal Monitor (<node2>) exists along with its <node1> counterparts.
7. For Service Managers and Internal Monitors of both app1 and app2 define a ‘Standard’ workshift (IMP- The workshifts are not defined by default for IMON/SM since the IMON is a feature exclusively for PCP).
8. Define a new concurrent manager of name Standard Manager2 with definitions same as that of a Standard Manager except that in the specialization rules form include a program, say, Active Users.
9. Exclude the Active User program from the specialization rules of the Standard Manager.
10. Stop and start the concurrent managers by executing the script adcmctl.sh.
11. Check whether the FNDSM and FNDIMON processes have started on both the nodes. It should be, otherwise it is indicative of miss-configuration.
12. Navigate to Concurrent > Manager > Administer form select the Standard Manager 2 and activate it.
13. After it has activated check from the back-end of the second node whether the desired FNDLIBR processes have been started. It should be, otherwise it is indicative of miss-configuration.
Regards,
Nitesh.