Wired Intelligent Edge

Hi

 

We use Aruba 2540 switches in our class rooms. When a student creates a loop (which occurs quite frequently) the switch disables the looped port thanks to loop-protect and automatically reenables the port after "auto-recovery-time" seconds. This works just fine.

 

However due to the high volume of network packets while the loop is active several other ports report "High collisions or drop rate" and shutdown as well. These ports then have to be reenabled manually. Is there a setting to reenable these ports automatically too?

 

Regards

That's interesting...could you share how the loop-protect context is configured on running configuration?

I observed high collisions or drop rate warning message too but not caused by loop-protect's detection of any other loop betweeb any other interfaces than the very first ones detected (I suspect because the loop-protect disable-timer was set too short - I rised it to 600 seconds - and this led the very first unfixed loop to kick in again too frequently even if its re-discovery was immediate).

We were able to solve the issue in the mean time. It seems the loop led to carrier errors on various other ports. Setting the fault-finder to log only did the trick.

 

The following command prevents the Aruba 2540 from disabling ports on

"High collision or drop rate" errors:

 

# fault-finder over-bandwidth sensitivity medium action warn

---

@Badzong wrote: The following command prevents the Aruba 2540 from disabling ports on "High collision or drop rate" errors:

 

# fault-finder over-bandwidth sensitivity medium action warn

---

Maybe the very same command with another action could do the trick:

 

fault-finder over-bandwidth sensitivity medium action warn-and-
disable

Did you fix using that action?

warn-and-disable was the configured behaviour that caused the problem in the first place. We are now measuring the used cabling to find out what is causing the dropped packets.

Greetings!

 

As part of your diagnostics, you can use the built-in cable tester to find out if there are any faulty connections from the switch in question. Use the following commands (note that this will cause a brief interruption of connectivity for each tested port):

 

switch# test cable-diagnostics 1-2 
This command will cause a loss of link on all tested ports and will take
several seconds per port to complete.  Use the 'show cable-diagnostics'
command to view the results.

Continue (y/n)?  y
switch# show cable-diagnostics 

 Cable Diagnostic Status - Copper Ports

       MDI    Cable       Cable Length or       
  Port Pair   Status      Distance to Fault     
  ---- ------ ----------- --------------------- 
  1    1-2    OK          2m                    
       3-6    OK          2m                    
       4-5    OK          2m                    
       7-8    OK          2m                    
  2    1-2    Open        0m                    
       3-6    Open        0m                    
       4-5    Open        0m                    
       7-8    Open        0m                    

In this example, port 1 is up and operating normally, and port 2 has no cable connected. If a cable is damaged, you may see Short or Open for one or more pairs, and the approximate distance from the switch port to the fault.

We are seeing the same issue as here, it looks like the config for 

 

"fault-finder over-bandwidth sensitivity medium action warn"

 

and 

 

"fault-finder over-bandwidth sensitivity medium action warn-and-disable"

 

are global though?

We want the options of having a single port doing read only or auto recovering. Is this an available feature?

 

TBB