WS-12-250DC Discarding, Learning and forwarding

[Steffan]

Hi Guys,

Hope you can help, I have a Netonix switch that seems to be every hour or so discarding, learning and forwarding on a port that is set to auto detect and has correctly identified the port as 1000mbps full duplex, powering a mimosa B5, this switch powers three B5's in total and the other two are fine, no TX/RX errors, no pause frames or anything odd that I can see going on, Cable seems fine and the B5 has been up and solid for over 3 days. I have no spanning tree links and none of the storm control properties have been enabled.

Any ideas welcome

[sirhc]

First you did not specify your Firmware version.
Upgrade to latest version which is v1.4.7rc11

The only cause of this:
1) The RJ45 connector is not firmly seated and is losing connection on one or more pins of Pins 1,2,3,4 which are required for 10,100,1G. Losing one of these pins will not result in the radio rebooting as it is getting power on other PINs.

2) Interference from FM at or around 100 MHz typically on FM tower sites.

In both cases above you will see the port link status change in the logs before the RSTP Blocking, Learning, Forwarding entries

3) A BPDU packet is getting back to the switch from another device connected to the switch possibly with a leaky VLAN.

Leaky VLANs are more often caused by using the Trunk Port option instead of defining detailed VLAN ID lists per port.

Another device is telling this switch to change a port status, this is is how RSTP works, the switches talk to each other and can tell each other to change port state. Read up on RSTP and how switches talk to each other and tell each other to do something.

You can turn RSTP OFF for that port and you will soon learn if there really is a loop.

I NEVER see RSTP issues at my WISP and I run VLANs and such and only use Netonix switches.

[Steffan]

Hi Sirhc

Thanks for your prompt advice, the firmware is 1.4.6 we've opted not to upgrade to 1.4.7rc11 as there are features still being worked on currently.

Should mention that the ports are not blocking they are discarding and there cannot be a loop on this part of our network as its relatively new.

I will get an engineer to site tomorrow to double check the cables etc., can you elaborate on the leaky vlan issue? we trunk the ports and add the vlans to the trunks i.e. vlan 20,10 then tag and untag or exclude them where they are needed etc.

Kind regards

Steff

[Julian]

I think pasting the relevant logfile entries might be a useful tool here, typical port behavior on link is below:

Dec 31 19:01:40 Port: link state changed to 'up' (1G) on port 4
Dec 31 19:01:40 STP: set port 4 to discarding
Dec 31 19:01:43 STP: set port 4 to learning
Dec 31 19:01:43 STP: set port 4 to forwarding

[sirhc]

You need to upgrade to v1.4.7rc11 to get support as we do not support older firmware than the current release.

The primary difference between v1.4.6 and v1.4.7rc11 is there are tons of bug fixes that were found in v1.4.6

We do not do things like other vendors.

If we finalized v1.4.7 tomorrow and then on Friday someone reports a bug in v1.4.7 then v1.4.8rc1 would be released.

You can read how RSTP works on the internet wiki and other publication, it is an IEEE standard 802.1w

We follow the IEEE standard.

https://en.wikipedia.org/wiki/Spanning_ ... e_Protocol

Leaky VLANs can allow BPDU packets to leak which can cause false positive.

Other switches or devices talking to our switch can instruct our switch to make changes to port status (Blocking, Leaning, Forwarding)

As I said you can disable RSTP on the port or all ports and see what happens, that will tell you real quick if there is a loop. If you have Loop protection on in our switch we will prevent the loop and there will be a more informative log entry telling you something.

[sirhc]

I think pasting the relevant logfile entries might be a useful tool here, typical port behavior on link is below:

Dec 31 19:01:40 Port: link state changed to 'up' (1G) on port 4
Dec 31 19:01:40 STP: set port 4 to discarding
Dec 31 19:01:43 STP: set port 4 to learning
Dec 31 19:01:43 STP: set port 4 to forwarding

If you see an entry in the log similar to the entry above in red where the physical link state changes that tells you the link is dropping and coming back on which cycles through the standard progression of RSTP which is blocking, learning, forwarding.

[Steffan]

Hi guys,

Just upgrading the firmware now....

Here is what I see in the logs:

Mar 7 04:54:24 STP: set port 11 to forwarding

Mar 7 06:38:00 STP: set port 11 to discarding

Mar 7 06:38:02 STP: set port 11 to learning

Mar 7 06:38:02 STP: set port 11 to forwarding

Mar 7 06:58:57 STP: set port 11 to discarding

Mar 7 06:58:57 STP: set port 11 to learning

Mar 7 06:58:57 STP: set port 11 to forwarding

Mar 7 07:08:21 STP: set port 11 to discarding

Mar 7 07:08:21 STP: set port 11 to learning

Mar 7 07:08:21 STP: set port 11 to forwarding

[Julian]

How did the firmware upgrade go?

Not sure what would be sending out BPDUs at ~20 minute intervals, I'm not the STP master around here though.. Seems like STP is deciding that there's a loop condition present?

a screen cap of your 'ports' tab might be nice; also what are the port statistics on the affected port saying?

[Steffan]

Hi Guys,

Thanks for all your help, I think I might have sorted it or it hasn't happened since at least.

It looks like a loose cable, not at the Netonix end but the far end, basically the Netonix switch was forwarding, discarding etc. because the Ethernet port on the far end (other side of the link) was dropping, loose clip on the network end got us chasing our tails for a couple of hours today.

Firmware upgrade went fine, very impressed with these switches so far.