We've recently discovered an issue with Resource Manager versions up to 6.1.0 wherein upgrades from 5.x don't expose required metrics for self-monitoring.
This issue has been documented as ZEN-29338 and will be fixed in a future maintenance release.
To work around this problem, customers can execute a migration script that corrects certain service definitions to expose the required metrics. To do so, follow these instructions:
Download the attached collectHBaseAndMariaStorageMetrics.py migration script and copy it to your Control Center master.
Execute the following commands to stage the migration script for execution:
mkdir /tmp/ZEN_29329 mv collectHBaseAndMariaStorageMetrics.py /tmp/ZEN_29329 serviced service shell --mount /tmp/ZEN_29329,/tmp/ZEN_29329 -i zope bash mv /tmp/ZEN_29329/collectHBaseAndMariaStorageMetrics.py /opt/zenoss/Products/ZenModel/migrate/ chown zenoss:zenoss /opt/zenoss/Products/ZenModel/migrate/collectHBaseAndMariaStorageMetrics.py su - zenoss
Verify that the migration script correctly lists in the output of the zenmigrate utility:
zenmigrate --list | grep CollectHBaseAndMariaStorageMetric
- The output should look like the following. If you receive a message like "UserWarning: poll_interval is ignored", you can safely ignore it.
storage = relstorage.storage.RelStorage(adapter, **relstoreParams) 200.0.0 CollectHBaseAndMariaStorageMetrics * Update MariaDB services to use supervisord.
Run the zenmigrate utility and then exit the container. There is no need to commit the container.
zenmigrate --dont-bump --step=CollectHBaseAndMariaStorageMetric exit exit
Before the changes can take effect, the altered services will need to be restarted.
serviced service restart Zenoss.resmgr