SG200-08 CPU hits 100% after running for one month

[cwork]

After running for about one month, my SG200-08 hit 100% CPU and pings increase from under 1ms to 300ms.   I purchsed the SG200-08 for home due to its support for IGMP snooping.  I have a TELUS Optik TV service at home which uses the Microsoft Mediaroom platform and multicast on the local LAN.   When the SG200-08 hits 100% CPU, my Cisco STBs start to exhibit multicast issues due to delayed or dropped IGMP messages.  

I recently upgraded the SG200-08 to firmware 1.0.6.2 hoping that it would fix the issue, but it hasn't helped.  

SG 200-08 8-Port Gigabit Smart Switch

System Uptime:  25 days, 16 hours, 37 mins 38 secs

Firmware Version: 1.0.6.2

PID VID: SLM2008T       V01

Smokeping latency graph of the SG200-08 ICMP response time.  On May 1st the CPU spiked to 100% and the laency increased to 300ms. 

The problem has been occuring since I installed the SG200-08 with the CPU spiking to 100% about once a month.  Rebooting the SG200 will clear the issue.

[Tom Watts]

Hi Clinton, this is pretty interesting.

Have you tried to disable features such as spanning tree or bonjour?

-Tom
Please mark answered for helpful posts

[cwork]

I had both Bonjour and Rapid STP enabled. I have disabled both for now and we'll see if the pegged CPU comes back in a month. 

I looked up the documenation for the Bonjour feature and it doesn't provide much value.  It just advertises some information about the switch and its http mgmt interface.   I'm not a big fan of the bonjour protocol, but if you use Apple devices or HP printers its a must.   I have found that most wireless APs don't implement IGMP snooping correctly and when enabled it breaks the Bonjour protocol.  The issue is an incorrect implementation of IGMP snooping that doesn't honor the well known multicast addresss in 224.0.0.x range that don't require IGMP join messages. 

[hanskbakke]

I see the same behaviour on my SG200-08 switch. After several weeks the CPU stays at full utilization until rebooted, and the latency skyrockets.

Have you reached any conclusions on bonjour/STP?

[PatrickB.]

Are there more people with that problem / is it confirmed? We recently bought a SG200-08p, let's see what happens in 30days....

Greets,

Patrick

[PatrickB.]

Just now the described problem showed up and a reboot fixed it.

I've now disabled Discovery - Bonjour and will wait another 20-30 days....

###

Uptime: 23 days, 0 hours, 25 mins 16 secs

CPU Utilization Report    

CPU Utilization For 5 seconds:     99%

CPU Utilization For 1 Minute:     99%

CPU Utilization For 5 Minutes:     99%

Firmware Version:     1.0.6.2

####

[PatrickB.]

In case that the problem occurs even after disabling Discovery - Bonjour, I will schedule an automatic reboot every Sunday night or something like that.

Just leaving that here:

root@XXXXX:~# cat X.X.X.X_reboot.sh

# Reboot X.X.X.X / SG200-08P / CPU Usage Bug

curl -c cookie -d 'uname=cisco&pwd2=XXXXXXXXXXXXXXX&language_selector=en-US&err_flag=0&err_msg=&passpage=nikola_main2.html&failpage=nikola_login.html&submit_flag=0' http://X.X.X.X/nikola_login.html

curl -b cookie -d 'v_1_2_1=Reboot&submit_flag=8&submit_target=AdminRebootStandalon.html&err_flag=0&err_msg=&dbgopt=0' http://X.X.X.X/AdminRebootStandalon.html/a1

[winston09]

I got 2 SG200-08 and have the similar problem.

Firmware version is the latest : 1.0.7.4

[hanskbakke]

I have not seen this issue after disabling bonjour discovery more than half a year ago. Since it could take a month between each occurrence it is hard to be absolutely sure that this was the cause, but it is worth mentioning.