Configuring the performance monitoring tool

Describes about configuring the performance monitoring tool in the IBM Storage Scale.

The performance monitoring tool, collector, sensors, and proxies are a part of the IBM Storage Scale distribution as detailed here:
  • The mmperfmon tool and the performance monitoring proxies are a part of the IBM Storage Scale core package gpfs.base, which is installed on all the nodes.
  • The sensors package gpfs.gss.pmsensors must be also installed on all the nodes with IBM Storage Scale if the performance monitoring is expected to provide information from these nodes.
  • The collector package gpfs.gss.pmcollector must be installed on all the nodes, which are expected to host the performance monitoring database. At least one node per 150 cluster nodes is recommended. Furthermore, it is recommended to not use NSD server nodes or quorum nodes for the installation. Using nodes with the IBM Storage Scale GUI makes most sense, since the IBM Storage Scale GUI queries the collector for its charts.

The performance monitoring tool uses the GPFS™ cluster daemon node names and network to communicate between nodes.

Note: The tool is supported on Linux® nodes only.

For information on the usage of ports for the performance monitoring tool, see the Firewall recommendations for Performance Monitoring tool.

The performance monitoring has a security feature to limit the connections of performance monitoring sensors to a list of known IBM Storage Scale node IPs. However, a user can modify the list of allowed IBM Storage Scale node IPs based on their requirements.

By default, this security feature is disabled, and can be enabled by modifying the ipfixiplist and ipfixiplistrefresh attributes in the /opt/IBM/zimon/ZIMonCollector.cfg configuration file. For more information, see the Configuring the collector