cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1229
Views
0
Helpful
2
Replies

Bandwidth Drop and Latency

STC_AliB
Level 1
Level 1

Hi Everyone,

 

I have been experiencing this issue for the past 2 weeks where a number of my clients reported slow connections speeds to from one of our remote sites back to main site. It was discovered that our Cisco 3850XS-S which is our layer 3 Switch on the site would not pass on traffic between 15-30Mbps to our routers. See image. (Note: After routers is the WAN link to other sites)Network_Layout.png

The issue of bandwidth and increased latency on the network did not occur like this previously. I have conducted file transfers over the WAN connection and confirmed that I can only obtain the same 15-30 Mbps speed to other sites. Further testing however has led to the following, our Layer 3 device refusing to process faster speeds. This was proven actioning a file transfer over the LAN only between the Packet analysis laptop directly connected to the Layer 3 Switch to a Virtual Machine hosted on a Vsphere Server (Link Between Server and Switch is 10Gbps cabling and configured for 10Gb). It seems like something has shaped the entire Layer 3 Switch to that speed. IPerf testing both LAN and WAN confirms the issue exists locally on LAN.

There is no QoS applied to the switch, nor has shaping been placed on any interfaces. All connections are as you would expect (Trunks between Switches, Access ports to Laptop, etc) and no configuration changes have been made to any of the network devices within the time period of these issues. 

The only noted issue is the version of IOS is old, 3.06.4 (rough guess) and does need to be upgraded, however I have yet to see any reported issues regarding latency and bandwidth speed on 3850XS models. CPU and Memory are well within acceptable limits.

 

Can anyone point me in the direction of where to continue troubleshooting from?? Or any Cisco tools that may assist in finding the cause? I am at this moment at a loss. Any assistance would be greatly accepted

2 Replies 2

Mark Elsen
Hall of Fame
Hall of Fame

 

 - For starters use a central syslog server. to collect all data from network devices. Examine the logs and watchout for irregular or abnormal messages.

M.



-- Each morning when I wake up and look into the mirror I always say ' Why am I so brilliant ? '
    When the mirror will then always repond to me with ' The only thing that exceeds your brilliance is your beauty! '

Apologies for the late reply.

 

We already have a syslog server pulling data from the network devices and there are no real errors that would be causing this. The only errors I am receiving are:

 

sw_matm-4-macflap_notif: Host 0050.5605.xxxx in vlan 124 is flapping between port Te1/0/1 and Po2

sw_matm-4-macflap_notif: Host 0050.5605.xxxx in vlan 121 is flapping between port Po2 and Te1/0/1

 

Otherwise there are no other error logs related to the devices in question.