na.hostperf manual page
Table of Contents

NAME

na.hostperf - host statistics

DESCRIPTION

na.hostperf is a proxy agent that reports statistics obtained from the kernel on the target system, including CPU, paging/swapping, disk and interface statistics. It gets data by querying the rpc.rstatd (1M) daemon on the target host through the rstat protocol.

A reporting interval is specified by the management application. If a reporting interval is not specified, the proxy agent sends reports every 30 seconds.

ATTRIBUTES

na.hostperf has one attribute group, data, that reports host data, i.e., perfmeter (1) information.

cpu% - CPU utilization percentage (gauge). The first time the agent reports this value, it reports the average since boot, that is, the total percentage the CPU was active since it was booted. After the first report, each report is the percentage the CPU was active since the previous report. intr - number of device interrupts since boot (counter) cswitch - number of context switches since boot (counter) avenrun_1 - average number of runnable processes over last minute (float)
avenrun_5 - average number of runnable processes over last five minutes (float)
avenrun_15 - average number of runnable processes over last 15 minutes (float)
disk - number of disk transfers since boot (counter) uptime - number of hundredths of seconds the host has been up (timeticks)
boottime - time the host was booted (unixtime) curtime - current time on host (unixtime) cpubusy - number of hundredths of seconds the CPU has been active since boot (counter)
pgin - number of pages read in from disk since boot (counter)
pgout - number of pages written to disk due to page faults since boot (counter)
pswpin - number of pages swapped in since boot (counter) pswpout - number of pages swapped out since boot (counter) ipkts - number of input packets on all interfaces since boot

(counter)
opkts - number of output packets on all network interfaces since boot (counter)
ierrs - number of input errors on all network interfaces since boot (counter)
oerrs - number of output errors on all network interfaces since boot (counter)
colls - number of collisions on all interfaces since boot (counter)
ocolls% - percentage of output collisions (gauge). The first time the agent reports this value, it reports the average since boot, that is, the total percentage of output collisions since the system was booted. After the first report, each report is the percentage that the CPU was active since the previous report.

OPTIONS

No options are accepted.

ERRORS

cannot create new subprocess; would exceed maximum: maximum The proxy agent cannot start a new request because creating a new subprocess would exceed the maximum number of proxy agent subprocesses allowed. Either terminate some requests and then resubmit the request or modify the value associated with the "na.hostperf.max-subprocs" keyword in /etc/opt/SUNWconn/snm/snm.conf for Solaris 2.x and in /etc/snm.conf for Solaris 1.x. If you change the "na.hostperf.max-subprocs" value, you must terminate all na.hostperf requests, kill all na.hostperf processes, and then restart the hostperf requests.

cannot get performance data
The rpc.rstatd (1M) daemon may not be running on the remote host, or the remote host may not be reachable.

NOTES

The way collision count is tallied depends on the Ethernet chip. The Intel (ie) chip counts every occurrence, while the Lance (le) chip counts collision on a per-packet basis. With the Lance chip, if a packet is successfully sent after the first retry, the collision count is one. If a packet is successfully sent after two or more retries, the chip returns a maximum count of two collisions. For example, if packet a had four retries and sometime later packet b had five retries, the cumulative collision count is four - two for each packet - not the real count of nine.

It is not possible for na.hostperf to report the system statistics per interface since the rpc.rstatd (1M) daemon does not break them out. Use na.hostif(8) to get perinterface statistics.


Table of Contents