CPU Idle information

[VF]

Hello everybody

I am using Cacti for monitoring Solaris and Linux boxes. On Solaris everything work fine but from Linux Cacti don't display CPU Idle information (nan).
When I run
$ snmpwalk -v3 hostname .1.3.6.1.4.1.2021.11 | grep -i idle
UCD-SNMP-MIB::ssCpuIdle.0 = INTEGER: 97
UCD-SNMP-MIB::ssCpuRawIdle.0 = Counter32: 219268660
In debug log I can see
04/18/2006 12:10:38 PM - CMDPHP: Poller[0] Host[49] DS[523] SNMP: v3: hostname, dsname: cpu_idle, oid: .1.3.6.1.4.1.2021.11.53.0, output: 215988327
04/18/2006 12:10:38 PM - CMDPHP: Poller[0] DEBUG: SQL Exec: "insert into poller_output (local_data_id,rrd_name,time,output) values (523,'cpu_idle','2006-04-18 12:10:01','215988327')"
When I try to run rrdtool info I get.
$ rrdtool info filename.rrd
filename = "filename.rrd"
rrd_version = "0001"
step = 300
last_update = 1145404802
ds[cpu_idle].type = "COUNTER"
ds[cpu_idle].minimal_heartbeat = 600
ds[cpu_idle].min = 0.0000000000e+00
ds[cpu_idle].max = 1.0000000000e+02
ds[cpu_idle].last_ds = "219331269"
ds[cpu_idle].value = 0.0000000000e+00
ds[cpu_idle].unknown_sec = 2
??.
May be any body have idea what is happed here? Looks that information lost in rrd file.
This is happened with cpu_idle only. RRD file changed every 5 min.

Thanks

[fmangeant]

Hi

can you post this :

Code: Select all

$ grep -c ^proc /proc/cpuinfo
4

[VF]

Hi,
I have 2 processors

$ grep -c ^proc /proc/cpuinfo
2

[fmangeant]

Then you need to modify the maximum value for this data source (with 2 CPU, idle will be between 0 and 200 %) :

Code: Select all

$ rrdtool tune filename.rrd -a cpu_idle:200

The other solution is to delete filename.rrd, as it will be created next time the poller runs.

[VF]

Thank you, this helps.
Why I don't need to do it for my 2, 4 and 8 ways SUN servers?

[fmangeant]

Net-SNMP doesn't work the same under Linux and the other Unixes.

There's a patch (available here : http://sourceforge.net/tracker/index.ph ... tid=312694), and it seems it has been integrated into Net-SNMP 5.3