This document shows how to perform a cleanup of data and correct issues when migrating from the ibmvm probe to ibmvm_rest.
Any UIM release
ibmvm version 2.52
1. Deactivate the ibmvm and ibmvm_rest probes
2. Delete all s_qos_defintion table entries for ibmvm_rest through a command similar to the following. **Note: All ibmvm_rest data will be lost for this QoS metric type**
delete from s_qos_definition where name like 'QOS_IBM_%';
3. Activate ibmvm_rest
4. The probe will publish a set of data that data_engine will start adding into s_qos_data.
5. It was found that there were two robot entries being populated and the data was being published to a different robot entry than what was expected. We used the following queries to validate:
select cs_key, name, ip, dedicated, orgin, mac from cm_computer_system where name = 'robot name'
select * from cm_device where dev_name = 'robot name'
6. From the above queries, it was found that one robot entry was being populated by the ibmvm probe and the other robot entry was being populated by the ibmvm_rest probe.
7. We removed the duplicate robot entries from the discovery_server probe utility by cs_key through the following method:
a. Navigate to discovery_server on the primary and press CTL P to bring up the probe utility.
b. In the combo-box (under the STOP sign), select "remove_master_devices_by_cskeys" and paste the cs_key from the query into the bottom text box labeled csKeys. Click the green triangle Play button. Perform this for one or more hosts. **Multiple cs_keys can be added at one time, comma separated**
c. Restart discovery_server
8. Lastly, we had to clear the niscache on both the robot, where ibmvm_rest was installed and on the robot that ibmvm_rest was remotely monitoring through the following:
a. Navigate to the controller probe on the primary and press CTL P to bring up the probe utility.
b. Click on the wrench icon and check the Expert Mode check-box on the bottom left.
c. In the combo-box (under the STOP sign), select the following
9. After performing these steps, all of the QoS metrics were correctly being populated under the expected robot.
10. Perform steps 5 - 7 on all robots having this same issue.
11. Clear the niscache on all of the robots that ibmvm_rest is remotely monitoring to have the QoS metrics correctly be populated under the robot.
Nis Cache Clean procedure:
At the Robot with the issue please:
Go to the Controller
Press CTRL +P
enable Expert mode
Select: _Nis_cache_clean ( execute the play green button)
Select : _reset_device_id_and_restart (execute the play green button)