clustervis - Man Page

visualize cpu and network performance on a cluster

Synopsis

clustervis [-H nodesfile] [-h host[,host...]] [-a archive[,archive...]] [-m max] [pmview options]

Description

clustervis displays three dimensional bar charts of CPU utilization and network traffic for one or most hosts in a cluster. An alternative two dimensional cluster performance monitoring tool is pmgcluster(1). clustervis is designed to provide a scalable overview of the performance of large clusters. Other tools such as pmchart(1) and pmgsys(1) provide drill-down details on a per-host basis. These tools may be launched by clicking on the purple base plane for a particular host and then selecting a tool from the launch menu in clustervis.

The -H, -h and -a arguments are all mutually exclusive and have the following semantics; if none of -H, -h or -a is given, and either the file /etc/nodes or /etc/ace/nodes exists, or the $PCP_CLUSTER_CONFIG environment variable is set, then use the named file as the set of hosts for live monitoring. If the default nodes file does not exist, the environment variable is not set and none of the three flags were given, an error is reported. Otherwise, if -H is given, then the set of hosts is given in nodesfile.

The -h flag specifies one or more (comma separated with no spaces) hosts for live monitoring and the -a flag specifies one or more archives for archive replay (comma separated).

The height of the CPU stack is proportional to the CPU utilization in each of the modes sys (red, executing in the kernel) and user (blue, executing user code). The network traffic stack is shown for each network interface as the packet rate in (light blue), out (orange) and errors (red). The hight of the network stack is modulated by the -m argument, with a default of 750 packets/second representing saturated network traffic.

clustervis generates a pmview(1) configuration file, and passes most command line options to pmview(1). Therefore, the command line options -A, -a, -C, -h, -n, -O, -p, -S, -t, -T, -Z and -z, and the user interface are described in the pmview(1) man page.

Files

$PCP_VAR_DIR/config/pmlogger/config.clustervis

A pmlogger(1) configuration file for clustervis metrics.

/usr/pcp/lib/pmview-args

Shell procedures for parsing pmview(1) command line options.

/etc/nodes

Default set of hosts in the cluster.

PCP Environment

The PCP_CLUSTER_CONFIG environment variable may be used to specify the default nodes file instead of using the -H flag. Each line in the file is a host name (or a comment starting with #). Other environment variables with the prefix PCP_ are used to parameterize the file and directory names used by PCP. On each installation, the file /etc/pcp.conf contains the local values for these variables. The $PCP_CONF variable may be used to specify an alternative configuration file, as described in pcp.conf(4).

See Also

dkvis(1), mpvis(1), nfsvis(1), pmcd(1), pmchart(1), pmlogger(1), pmview(1), pcp.conf(4) and pcp.env(4).

The CPU view for pmchart(1).

Info

Performance Co-Pilot