CPU Switch Catalyst 6500 VSS

[ecolmenaresb1]

hello experts

i need your help

I administer a catalyst switch 6500 virtualized in vss,  I need to know the cpu levels of these switch, Because it has a total consumption of 80%, and interruption 16%  For periods of time of 20 or 30 minutes and then down to 60%, but this behavior is constant, the company is in the study of replacement switch, but I need to indicate in which percentage of cpu this switch can work optimally

I await your answers, thank you very much

[Mark Malone]

Hi that's pretty high if its staying at that level even at constant 60%, whats the proc output showing you as the main cause for this high cpu ?

run this when its high

sh proc cpu sorted | ex 00.0

[ecolmenaresb1 · Accepted solution]

hello

Attach command on the vss switch

CPU utilization for five seconds: 89%/25%; one minute: 83%; five minutes: 79%
 PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process
 532    38996600  55663348        700 18.87% 17.39% 15.71%   0 DHCPD Receive    
 303   5217792801975394339        264  9.75%  9.23%  8.68%   0 IP Input         
 481   9674538724250513307          0  8.87%  4.69%  4.48%   0 Port manager per
 286   604888992 763918829        791  5.11%  4.08%  4.08%   0 Earl NDE Task    
 538   8655468602167220016          0  4.23%  6.79%  6.93%   0 SNMP ENGINE      
 297   257566680 190291807       1353  2.23%  2.20%  2.17%   0 CDP Protocol     
   3    10426092 175016736         59  0.79%  0.73%  0.72%   0 LLDP Protocol    
 367   130877704  51904435       2521  0.79%  0.84%  0.84%   0 CEF: IPv4 proces
  23   4992097923556810342          0  0.79%  2.19%  2.06%   0 IPC Seat Manager
  59    734871282098971831         35  0.63%  0.36%  0.35%   0 VSIBC process    
  85    62082612 125072015        496  0.47%  0.39%  0.39%   0 ARP HA           
 142          64        88        727  0.47%  0.06%  0.01%   1 Virtual Exec     
 119    64308388  51281726       1254  0.39%  0.35%  0.34%   0 IPC LC Message H
 305    74515368 598890534        124  0.31%  0.38%  0.39%   0 ADJ resolve proc
 259    10693012 160818121         66  0.23%  0.22%  0.23%   0 cpf_process_tpQ  
  50    16860144  30408861        554  0.23%  0.31%  0.31%   0 Per-Second Jobs  
 544    10463088 458227631         22  0.23%  0.25%  0.23%   0 EIGRP-IPv4 Hello
 323     14340962988403985          0  0.23%  0.24%  0.23%   0 Ethernet Msec Ti
  63    293535481483051535         19  0.15%  0.18%  0.17%   0 Net Input        
 536   106743640 862439120        123  0.15%  0.31%  0.34%   0 IP SNMP          
  20     6754684  23766476        284  0.07%  0.07%  0.07%   0 IPC Periodic Tim
 480     2331256 775147282          3  0.07%  0.08%  0.07%   0 PM Callback      
 409     1865596 732602509          2  0.07%  0.06%  0.07%   0 RADIUS           
 448    17654580 107475489        164  0.07%  0.06%  0.07%   0 RPC online_diag_
 128    19363256   6869918       2818  0.07%  0.08%  0.07%   0 Compute load avg
 288    27000228  72392700        372  0.07%  0.12%  0.12%   0 esw_vlan_stat_pr
 537    45179880 435838478        103  0.07%  0.13%  0.14%   0 PDU DISPATCHER   
 391     9717984  11896836        816  0.07%  0.04%  0.05%   0 HIDDEN VLAN Proc

[Mark Malone]

so you could have a culprit dhcp device causing that high DHCPD process or do you have interfaces with multiple helper addresses set per interface ?

There is a way of capturing traffic going to the cpu to see what the cause is if its a suspect device , if your not familiar it you could ask TAC to run it

https://supportforums.cisco.com/document/59956/troubleshooting-netdr-capture-sup7206500

what version of code are running currently on the VSS ?

This is also very good doc on things to check when seeing high cpu in 6500 to make sure you don't have set that can cause issues too

https://supportforums.cisco.com/document/59926/troubleshooting-high-cpu-6500-sup720

are you seeing anything in logs as well ?

[ecolmenaresb1]

ok thank you very much

Now what are the appropriate cpu levels for a 6500 switch

Ej:

0-60% Normal

60-80 % warning

80-100 % critical

Please some cisco document that indicates this since I need to back up with the client

[Utkarsh Dewangan]

Hi,

It seems you have a large amount of DHCP traffic running through the switch.

DHCPD could go high when multiple SVI have IP helper addresses configured in them.

As all the DHCP broadcasts are handled by CPU.

check with following if you are seeing a frequent increase.

show ip DHCP relay statistics.

if feasable check with removing relay addresses

also the IP input is high pointing that traffic is being punted to cpu.

you can check with taking a cpu capture, eg netdr.

https://supportforums.cisco.com/document/59956/troubleshooting-netdr-capture-sup7206500

see if you see any specific host sending excessive broadcasts.

PS: sorry about the endorsement, clicked by mistake