This is a known issue on R4.0 that whenever the uptime of CM server exceeds 500+ days, a CPU PERF minor alarm comes in the IA770 Messaging.
The uptime of this system was 1320. Also, the Date of first CPUPERF warning/alarm is when uptime equals b/w 500-514 days.
To verify this, run two commands "sar" and "top". "top" command shows sar command miscalculates the system resources' usage and raises the alarm while you can see actual resources usage in "top" command. By the time this issue was identified, development on this CM release was already stopped.
Please note that this is fake alarm. The CPU utilization in system is normal if top command shows so. System has enough CPU resources.
init@> uptime
12:28:42 up 1320 days, 21:50, 2 users, load average:
1.74, 1.34, 1.20
init@>
% sar 5 5
Linux 2.6.11-AV18 () 04/28/11
12:44:55
CPU %user %nice %system %iowait %idle >>>> 0% idle
12:45:00 all
59.32 0.00 38.98 1.69 0.00
12:45:05 all 49.19
0.00 50.81 0.00 0.00
12:45:10 all 48.03 0.00
50.88 1.10 0.00
12:45:15 all 44.19 0.00 55.81
0.00 0.00
top - 12:45:24 up 1320 days, 22:07, 3 users, load average:
0.92, 0.75, 0.85
Tasks: 266 total, 1 running, 265 sleeping, 0 stopped, 0
zombie
Cpu(s): 7.2% us, 6.2% sy, 0.0% ni, 86.3% id, 0.0% wa, 0.0% hi, 0.3% si >>> 86% idle
Mem: 515420k total, 501476k used, 13944k free,
84692k buffers
Swap: 530136k total, 184k used, 529952k free, 115892k
cached
