04-04-2014 07:43 AM - edited 03-07-2019 06:59 PM
Hi all,
I have planned to change all oldest access switches with 2960X. On a test one, monitored by PRTG, CPU is always around 40% as the old 3560's are 8% CPU usage.
Running the EX4 release. Does anyone have the same values ? This problem seems to be recurrent on some Cisco switch a since years....
Nothing in EX5 release notes about that.
Thanks
FBL
04-04-2014 03:33 PM
Let me brutally frank with you, the current IOS for 2960X/XR is, for a lack of a better word, risky.
EX3 is considered by TAC to be "stable" (suuuuuuuuuuuuuuuure). EX4 has known CPU bugs.
In my book, you'll need to wait before the main line release of the 15.X comes out. I seriously wouldn't recommend anyone using EX-series IOS for production if the main line gets release.
06-21-2016 06:25 AM
Running the latest & greatest with realistic expectations. This IOS release date is 17-FEB-2016 and the CPU utilization is still pretty high. This switch has nothing connected except for console and power. All interfaces are shutdown and all copper interfaces are configured "power inline never."
Before I made changes, the switch was showing 37% CPU utilization, mostly from the Hulc LED Process. This switch has no stack modules and is running the LAN base image.
Switch#show version
Cisco IOS Software, C2960X Software (C2960X-UNIVERSALK9-M), Version 15.2(2)E4, RELEASE SOFTWARE (fc2)
~output omitted~
ROM: Bootstrap program is C2960X boot loader
BOOTLDR: C2960X Boot Loader (C2960X-HBOOT-M) Version 15.2(3r)E1, RELEASE SOFTWARE (fc1)
Switch uptime is 45 minutes
System returned to ROM by power-on
System restarted at 12:36:58 UTC Tue Jun 21 2016
System image file is "flash:/c2960x-universalk9-mz.152-2.E4/c2960x-universalk9-mz.152-2.E4.bin"
Last reload reason: Reload command
~output omitted~
cisco WS-C2960X-48LPS-L (APM86XXX) processor (revision K0) with 524288K bytes of memory.
~output omitted~
Last reset from power-on
1 Virtual Ethernet interface
1 FastEthernet interface
52 Gigabit Ethernet interfaces
The password-recovery mechanism is enabled.
512K bytes of flash-simulated non-volatile configuration memory.
~output omitted~
Motherboard assembly number : 73-16692-04
Power supply part number : 341-0528-01
~output omitted~
Model revision number : K0
Motherboard revision number : A0
Model number : WS-C2960X-48LPS-L
Daughterboard assembly number : 73-14200-03
~output omitted~
Top Assembly Part Number : 68-100470-01
Top Assembly Revision Number : A0
Version ID : V03
CLEI Code Number : CMMLP00ARC
Daughterboard revision number : A0
Hardware Board Revision Number : 0x18
Switch Ports Model SW Version SW Image
------ ----- ----- ---------- ----------
* 1 52 WS-C2960X-48LPS-L 15.2(2)E4 C2960X-UNIVERSALK9-M
Configuration register is 0xF
Switch#show processes cpu sorted 5min
CPU utilization for five seconds: 29%/0%; one minute: 29%; five minutes: 30%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
160 604129 67145 8997 15.06% 14.99% 15.14% 0 Hulc LED Process
171 8252 552 14949 0.23% 0.25% 0.23% 0 HQM Stack Proces
6 4247 462 9192 0.00% 0.10% 0.11% 0 Check heaps
140 3414 1132 3015 0.00% 0.00% 0.04% 0 Exec
386 2145 106 20235 0.00% 0.00% 0.01% 0 hulc running con
21 24 545 44 0.05% 0.00% 0.00% 0 IPC Event Notifi
117 51 2685 18 0.00% 0.00% 0.00% 0 Hulc ILP Alchemy
172 1274 1088 1170 0.00% 0.00% 0.00% 0 HRPC qos request
8 0 1 0 0.00% 0.00% 0.00% 0 DiscardQ Backgro
7 0 1 0 0.00% 0.00% 0.00% 0 Pool Manager
10 7 280 25 0.00% 0.00% 0.00% 0 WATCH_AFS
06-21-2016 03:26 PM
Peter,
Are there any PoE &/or Dot1X ports enabled?
06-22-2016 07:01 AM
I also revisited the Cisco Catalyst 2960-X Series Switches Data Sheet (http://www.cisco.com/c/en/us/products/collateral/switches/catalyst-2960-x-series-switches/data_sheet_c78-728232.html) and the CPU is a APM86392 600MHz dual core. So, I googled that and got the data sheet on the CPU: http://c1170156.r56.cf3.rackcdn.com/UK_AMC_APM86392-SGA600T_DS.pdf. According to the CPU datasheet, it could be faster.
To answer your question on that switch, all interfaces are shutdown and all copper interfaces are configured "power inline never."
I have another switch running the same code part & it's in a pod I'm configuring for a remote site. So, that one has dot1x, QoS, but again, no PoE and all SFP ports are disabled. Here's the interesting parts of the configuration:
no ip domain-lookup
ip domain-name [mycompany].com
vtp mode transparent
udld aggressive
mls qos map cos-dscp 0 8 16 24 32 46 48 56
mls qos
spanning-tree mode rapid-pvst
spanning-tree loopguard default
spanning-tree portfast bpduguard default
spanning-tree extend system-id
auto qos srnd4
errdisable recovery cause udld
errdisable recovery cause bpduguard
errdisable recovery interval 60
no cdp run
interface Port-channel1
description Uplink_SW1
switchport trunk allowed vlan [3 VLANs]
switchport mode trunk
switchport nonegotiate
#Uplink switchport configuration:
interface GigabitEthernet1/0/1
description Uplink_SW1_Gi1/0/46
switchport trunk allowed vlan [3 VLANs]
switchport mode trunk
switchport nonegotiate
power inline never
channel-group 1 mode active
#Access switchport configuration:
interface GigabitEthernet1/0/2
switchport access vlan [workstation VLAN]
switchport mode access
switchport nonegotiate
switchport port-security maximum 2
switchport port-security violation restrict
switchport port-security aging time 5
switchport port-security aging type inactivity
switchport port-security
power inline never
speed auto 100
storm-control broadcast level 10.00
storm-control multicast level 10.00
And here's the CPU utilization (didn't make much of a difference):
SW3#sho proc cpu sort 5min
CPU utilization for five seconds: 29%/0%; one minute: 29%; five minutes: 29%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
160 11096203 1587145 6991 15.00% 14.98% 14.99% 0 Hulc LED Process
166 2697 576 4682 0.00% 0.00% 0.12% 0 Exec
171 90008 12697 7088 0.11% 0.11% 0.11% 0 HQM Stack Proces
6 101067 10982 9202 0.00% 0.09% 0.10% 0 Check heaps
172 37016 25386 1458 0.05% 0.04% 0.05% 0 HRPC qos request
389 1479 66 22409 0.00% 0.00% 0.01% 0 hulc running con
However, this pod of equipment (2 x 1921 routers & 3 2960X switches) are not under load. I will be monitoring this equipment after it's deployed and plan to post the results.
06-22-2016 02:09 PM
errdisable recovery cause bpduguard
It won't help the issue, but I'd take this off. BPDU Guard is a network admin's friend. If someone's plugged an illegal switch in the network I sure would like to know about it. Enabling a "recovery" is like sweeping the issue under the carpet.
06-22-2016 07:17 PM
We played with some of the errdisable settings which will be removed when this switch and others move to production. Most remote sites won't need fiber, so we're also getting rid of all udld commands. I have bpduguard enabled on pretty much all my production switch access ports.
06-22-2016 04:32 PM
Hi,
This is an expected behavior on the 2k and 3k switches where the Hulc led process will be running between 15-30% even on a unit with no load on it.
Regards,
04-09-2014 05:51 AM
Hi,
This is an expected behavior and there is no workaround.
The "Hulc LED" process does following tasks:
- Check Link status on every port
- If the switch supports POE, it checks to see if there is a Power Device (PD)
detected
- Check the status of the transceiver
- Update Fan status
- Set Main LED and ports LEDs
- Update both Power Supplies and RPS
- Check on system temperature status
Hope this helps.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide