Originally posted by: thloeber

Following the install instructions provided from the IBM Knowledge Center I managed to get a Platform Conductor Master Host installed on ppc64 RHEL 7.1. Moving on the the configuration and entitlement steps I've run into problem.  I ran the install as root (sudo -i) which ran without obvious errors.  The next step in the install said to run $EGO_TOP/profile.platform which I did as root as well.  These are those instructions http://www.ibm.com/support/knowledgecenter/SSVH2B_1.1.0/install/install_linux_management.dita

These instructions to configure the Platform Conductor is where I have the issue to set the entitlement. The egoconfig commands to run to join the host and set the entitlement do not run (cannot be found) as root. I also tried running those same commands as the clusteradmin which is egoadmin and these could not be found either. The product is installed at /opt/ibm/platform.  As who do this commands need to be run? I believe there is a step missing  as far as setting up the environment under which the platform conductor needs to run. I assume these need to run as egoadmin but the path variable is not set properly and I imagine there are other environment variables that need to be set and exported in the .bash_profile as well. Are there steps missing from these instructions?

http://www.ibm.com/support/knowledgecenter/SSVH2B_1.1.0/install/install_set_entitlement.dita

Originally posted by: thloeber

After opening a PMR I did get resolution on this. The configuration steps as explained at http://www.ibm.com/support/knowledgecenter/SSVH2B_1.1.0/install/install_set_entitlement.dita omit  a step where the /opt/ibm/platform/profile.platform needs to be chmod +x as do the other for scripts called in that profile. This also needs to be run as the CLUSTERADMIN that was specified during the installation process, in my case egoadmin.  The CLUSTERADMIN also needs to be granted root permissions in order to successfully execute the egosg ego start command. This is explained further into the document in the Managing Your Cluster section. By default the cluster startup can only be run by root, which would have to have the environment source as well.

Originally posted by: 6K3S_John_Nguyen

Hi Thomas,

Thank you for your posts.

With regard to the issue you encountered, I think you might have missed the "." when running the command to source the environment:

[user1@host1 opt]
$ . $EGO_TOP/profile.platform

(found in step 8 in http://www.ibm.com/support/knowledgecenter/SSVH2B_1.1.0/install/install_linux_management.dita)

This command will automatically setup your current shell environment with the proper paths and environment variables, allowing you to run commands that start with "egosh" or "egoconfig".  So you would not need to change the file permission of profile.platform using 'chmod +x'.

However, if the user logs out of the shell and then logs back in, they will be required to re-run the command above to source the environment again.  Some of our users have added this command to their CLUSTERADMIN user's bash profile to eliminate this.

Originally posted by: thloeber

Thank you for the post reply.  I have added the this command to the CLUSTERADMIN bash profile.  I am more of a system administrator then an actual Spark user. I am tasked with getting this up and running for other memebers on my team.  I see there is an ego script in /etc/init.d as well.  But the Managing Your Cluster section does not mention starting and stopping the ego cluster using systemctl . Is that not recommended?

Originally posted by: 6K3S_John_Nguyen

Hi Thomas,

The "ego" script you see in /etc/init.d was likely created when you ran egosetrc.sh, which allows for automatic startup of Conductor's daemons/services when the host is booted up. (https://www.ibm.com/support/knowledgecenter/SSVH2B_1.1.0/reference_sym/egosetrc.8.dita?lang=en).

We don't use 'systemctl' to manually start and stop the daemons/services for Conductor.

To start and stop services manually, we just use use the 'egosh' commands as you've already seen in the "Managing Your Cluster" section (https://www.ibm.com/support/knowledgecenter/SSVH2B_1.1.0/manage_cluster/manage_cluster.dita?lang=en).