Applies to: All versions of Centrify DirectControl 5.1.x on Solaris platforms.
When running the top command on a Solaris host with limited RAM, why does it reports high cpu on a machine which does not have adequate CPU power?
Note: The following steps were attempted and have been confirmed to have no affect on resolving the issue:
1) nscd was turned on and it did not alleviate the high CPU usage.
2) In /etc/centrifydc/centrifydc.conf, the adclient.cache.expire parameter was changed form 600 to 3600 and adreload was executed.
load averages: 1.27, 0.73, 0.62 09:45:45
141 processes: 134 sleeping, 6 zombie, 1 on cpu
CPU states: 84.6% idle, 10.7% user, 4.7% kernel, 0.0% iowait, 0.0% swap
Memory: 2048M real, 634M free, 659M swap in use, 16G swap free
PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
24241 root 8 59 0 29M 23M sleep 3:13 17.69% adclient
19129 reissmr 1 59 0 75M 67M sleep 17:07 5.48% xzt
470 root 18 59 0 8680K 4952K sleep 1:07 4.55% syslogd
12966 root 1 59 0 4352K 1912K cpu/1 0:00 0.10% top
25179 root 1 59 0 4544K 1608K sleep 4:01 0.07% top
When running the top binary on machines with low RAM, the top command is looking for the resource limit of number of open files allowed. It can report inaccurate results.
It is recommended to run the built-in Solaris command prstat command instead.
Note: The below link compares both tools. Its provided as a courtesy. Centrify takes no responsibility for the content or availability at a later date and time.