Spanning-Tree CPU utilization on a 2960

zstamm · ‎08-21-2018

I have a 2960 that has high CPU utilization. I know the "Hulc LED Process" processor usage is caused by a glitch, but what about the spanning tree CPU usage? There are 176 VLANs running on this switch, and it is running Rapid PVST+. Is that typical CPU usage for this kind of switch running that many instances of spanning-tree?

The CPU utilization is as follows:

XXXXXXXX#sh proc cpu sorted | ex 0.00
CPU utilization for five seconds: 81%/2%; one minute: 77%; five minutes: 77%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
220 1041510985 88802676 11728 41.27% 35.86% 35.65% 0 Spanning Tree
162 627025960 43046544 14566 16.86% 19.63% 20.17% 0 Hulc LED Process
357 13330467 6354260 2097 3.98% 1.07% 0.65% 0 SNMP ENGINE
131 27388549 2603493 10519 0.93% 1.00% 1.00% 0 hpm counter proc
7 8194435 649655 12613 0.35% 0.24% 0.25% 0 Check heaps
173 8247868 513962 16047 0.29% 0.27% 0.28% 0 HQM Stack Proces
93 5424095 61881041 87 0.17% 0.14% 0.16% 0 RedEarth Rx Mana
194 3267192 2240256 1458 0.17% 0.11% 0.11% 0 CDP Protocol
236 3685301 2534533 1454 0.11% 0.12% 0.12% 0 PI MATM Aging Pr
207 1054175 6843668 154 0.11% 0.13% 0.08% 0 IP Input
355 3839010 12702559 302 0.11% 0.27% 0.22% 0 IP SNMP

Reza Sharifi · ‎08-21-2018

There are 176 VLANs running on this switch, and it is running Rapid PVST+

That is a lot of vlans running on an access switch with a maximum of 48 ports. I would consider deleting some of these vlans if you are not using them.

A Layer 2 spanning-tree instance runs for every VLAN configured on a Layer 2 switch by the per-VLAN spanning-tree (PVST) feature. The CPU time utilized by spanning tree varies depending upon the number of spanning-tree instances and the number of active interfaces. The more instances and the more active interfaces, the greater the CPU utilization.

Link:

https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst3750/software/troubleshooting/cpu_util.html#38092

HTH

Joseph W. Doherty · ‎08-22-2018

"The CPU time utilized by spanning tree varies depending upon the number of spanning-tree instances and the number of active interfaces. The more instances and the more active interfaces, the greater the CPU utilization."

Just wanted to add, to what Reza notes, that what I've bolded might be mitigated by using port fast. Are you using it on edge interfaces?

Also I might add, pruning VLANs will help mitigate this too.

Lastly, you might look at MST, as I doubt you really need so many PVST instances.

Leo Laohoo · ‎08-21-2018

@zstamm wrote:

There are 176 VLANs running on this switch, and it is running Rapid PVST+.

I'm with Reza. That's one badly designed network right there. All you need is one of those links flapping and this switch is dead.