CPU load on VCS?

[RICHARD MESSINGER]

have a customer that is starting to monitor performance and health on their VCS controller with Solarwinds.

They got an alert this morning when the CPU#1 and #2 reported 100%.

1) is this "normal"

or

2) is Solarwinds not checking the correct MIB values?

VCS is running 7.1 code

There do not seem to be any messages in TMS or on the VCS itself.

Thoughts?

[William Bell]

I haven't tried to monitor a VCS via a 3rd party SNMP management system. My understanding is that Tandberg was more into using a XML-based API for element monitoring. That said, I believe the VCS does support the basic MIB objects for *nix and that is probably what you are using. It may show up as 100% because the system is rolling up the idle process in the statistic. Just a guess. I know I have seen tools show 100% CPU utilization on *nix systems when OS (or API) tools would show that most of that was idle time.

I am sure someone can jump in and give you more specific guidance.

HTH.

-Bill

[rasimyigit]

Make restart and check it again

Sent from Cisco Technical Support iPhone App

[Martin Koch]

It might be interesting to know which oids are queried by your management software.

Like Bill said the standard linux mibs seem to be supported.

I am not aware if its possible to get the cpu load via the vcs-xml-api

If you want to check whats going on its worth logging into your VCS as root and check

with some commands what the actual load is.

If you execute: top -bn1

you will see something like:

~ # top -bn1 -n1
top - 11:36:01 up 10 days,  5:21,  3 users,  load average: 0.27, 0.54, 0.60
Tasks: 146 total,   1 running, 145 sleeping,   0 stopped,   0 zombie
Cpu(s):  6.4%us,  1.6%sy,  0.3%ni, 91.4%id,  0.1%wa,  0.0%hi,  0.2%si,  0.0%st
Mem:   4042548k total,  3991536k used,    51012k free,     6536k buffers
Swap:  9775516k total,    29584k used,  9745932k free,   462696k cached
  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND            
 5206 root      20   0 1362m 1.0g  12m S   22 25.6   1554:16 app                
 3118 root      20   0 89132  39m 2128 S    4  1.0 224:49.33 syslog-ng          
 6679 root      20   0  763m 566m 3480 S    4 14.3 164:24.21 java               
    1 root      20   0  3888  604  520 S    0  0.0   0:04.63 init               
    2 root      20   0     0    0    0 S    0  0.0   0:00.22 kthreadd           
    3 root      20   0     0    0    0 S    0  0.0   1:09.13 ksoftirqd/0        
    6 root      RT   0     0    0    0 S    0  0.0   0:01.88 migration/0        
    7 root      RT   0     0    0    0 S    0  0.0   0:01.65 migration/1        
...

you can also run top without the options to get some continous output, you can stop it with q or ctrl-c.

By pressing 1 in you can also show the usage per cpu.

If something goes wong its most likely to higl load of java or app. And if you really have 100%, no that

should not be normal.

There was a bug with the leap second but x7.1 should not be affected. I am not aware if there

if any specific bug you hit, maybe searching the bug toolkit can give you more info.

If you see some high load / cpu usage a reboot will most likely fix it.

If not (and anyhow), x7.2.1 is the latest code available.