CPU Usage for Solaris

[stigmata]

Yes, problem solved

One other question - "Wait" is this the I/O Wait? I need the values for the I/O wait.
Uuh, other question, can you explain "ticks/s"? it would be very helpful for me.

thx - Regards,
Oliver

[o_dupuis]

Hi,

Yes indeed Wait is short for I/O Wait.

About the timeticks, it means that you dont get a percentage but a number of ticks per seconds. This is the way the OS gives it.
To get a percentage, just make a CDEF that makes the division of the measurement (WAIT, IDLE, KERNEL or USER) by the <total all data srouces nodup>, and that then multiplies it by 100. There you get your percentage.


Olivier

[kiw0r]

Just a note to everyone doing this- not all of these CPU values are present on a given architecture. For example, Linux 2.6 will have "nice" CPU, but Solaris does not. And, not all versions of net-snmp have support for them.

I have just done some Template for FreeBSD.
You can use them if you like..
Note to Cacti developpers, can these template (solaris, linux, freebsd, etc..) be added into next cacti ? It should good to not look around everything there ?

[adrianmarsh]

Anyone know why some of my net-snmp'd solaris boxes would produce a graph like this? It seems to skip a beat every now and then. Other solaris boxes are ok, but I've at least 2 like this. Up time stats on the devices page says 100%

[pat_ra2001]

I'd like to start snmp service on sun solaris. Please help me.

[thehemi]

Hi dgamble,

thx for your great CPU-Script. I use the one for the 2.6 linux kernel. It works great but one thing, I get a "nan" at the "Wait" value (but this is very importend for me ).

What can I do?

Regards,
Oliver

I believe that net-snmp returns that object only as of version 5.2, which is probably the issue.

Sorry for bringing back an old thread, but I'm having this issue now.

net-snmp-5.0.9-2.90.1
net-snmp-5.1.2-11

The 5.0.x system does NOT return the "Wait".
The 5.1.x system returns the "Wait" and others.

But upgrading all of the necessary packages to get net-snmp-5.1.2
installed on a RHEL 3 box is a pain in the butt. I hate dependencies.

[rcl_1973]

On my Solaris 8 server I have snmpd working fine, but I need other oids.

When I run "snmpwalk" from outside (from a linux with net-snmp) I can only get values from this categories:

IF-MIB
IP-MIB
RFC1213-MIB
SNMPv2-MIB
TCP-MIB
UDP-MIB

I need check CPU, MEM, etc (they are inside UCD-SNMP-MIB) but I don't know how to install it on Solaris 8 (without installing net-snmp for solaris).

Thanks

[sys0p]

It seems that cpu_nice doesnt show up on the graph.

In the log file I see :

08/17/2007 12:40:01 AM - CACTID: Poller[0] Host[2] DS[135] SNMP: v2: 192.168.0.11, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 7320
08/17/2007 12:40:01 AM - CACTID: Poller[0] Host[3] DS[19] SNMP: v2: 192.168.0.12, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 409
08/17/2007 12:40:01 AM - CACTID: Poller[0] Host[4] DS[29] SNMP: v2: 192.168.0.13, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1146
08/17/2007 12:40:02 AM - CACTID: Poller[0] Host[5] DS[40] SNMP: v2: 192.168.0.1, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 87
08/17/2007 12:40:02 AM - CACTID: Poller[0] Host[6] DS[52] SNMP: v2: 192.168.0.101, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1553
08/17/2007 12:40:06 AM - CACTID: Poller[0] Host[8] DS[73] SNMP: v2: 192.168.0.102, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1518
08/17/2007 12:45:01 AM - CACTID: Poller[0] Host[2] DS[135] SNMP: v2: 192.168.0.11, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 7320
08/17/2007 12:45:02 AM - CACTID: Poller[0] Host[3] DS[19] SNMP: v2: 192.168.0.12, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 409
08/17/2007 12:45:02 AM - CACTID: Poller[0] Host[4] DS[29] SNMP: v2: 192.168.0.13, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1146
08/17/2007 12:45:02 AM - CACTID: Poller[0] Host[5] DS[40] SNMP: v2: 192.168.0.1, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 87
08/17/2007 12:45:02 AM - CACTID: Poller[0] Host[6] DS[52] SNMP: v2: 192.168.0.101, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1553
08/17/2007 12:45:06 AM - CACTID: Poller[0] Host[8] DS[73] SNMP: v2: 192.168.0.102, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1518
08/17/2007 12:50:02 AM - CACTID: Poller[0] Host[2] DS[135] SNMP: v2: 192.168.0.11, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 7320
08/17/2007 12:50:02 AM - CACTID: Poller[0] Host[3] DS[19] SNMP: v2: 192.168.0.12, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 409
08/17/2007 12:50:02 AM - CACTID: Poller[0] Host[4] DS[29] SNMP: v2: 192.168.0.13, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1146
08/17/2007 12:50:02 AM - CACTID: Poller[0] Host[5] DS[40] SNMP: v2: 192.168.0.1, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 87
08/17/2007 12:50:02 AM - CACTID: Poller[0] Host[6] DS[52] SNMP: v2: 192.168.0.101, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1553
08/17/2007 12:50:06 AM - CACTID: Poller[0] Host[8] DS[73] SNMP: v2: 192.168.0.102, dsname: cpu_nice, oid: .1.3.6.1.4.1.2021.11.51.0, value: 1518

The rrd file is obviously created.. but when looking at the graph the :
Nice CPU Data Source
The data source to use for this graph item. is set to None

Even if I choose the right rrd from the list and save it , next time I open it it still shows None.

Can someone please explain this ?

[alexdu]

hi all, if there have any templates for windows server 2003? Plz help to offer one or give some guide to solve this issue, thanks a lot!

[cerebrumaugite]

Hi dgamble,

thx for your great CPU-Script. I use the one for the 2.6 linux kernel. It works great but one thing, I get a "nan" at the "Wait" value (but this is very importend for me ).

What can I do?

Regards,
Oliver

I believe that net-snmp returns that object only as of version 5.2, which is probably the issue.

How can i reduce my electricity cost? I am working out on a project to make the electricity bill for computing disappear? I have decided to try to find out.I also spent time with the devices - autosuspending as many devices as possible. The major applications – such as Apache, sendmail, NTP, etc – were tuned for CPU efficiency, and any necessary applications and hardware (eg the backlight and WiFi) were disabled. The networking was simplified to reduce consumption to 8W for one ADSL router. The upshot of this part of the story is that it all worked and looking for more option