Request Feedback v1.1.0 - PLEASE

[sirhc]

Ha!

Totally need to quit using my phone for forum replies!

Any update on the stp issue Chris, Rory?

I apologize, Rory has been out of the office since Monday with a family emergency and my partner Rusty died Sunday so I have not been able to LAB it but Rory would have been better to LAB it then me as I am very rusty which is why I pay him.

Sadly Eric is also off today since it is his birthday (tell him he is an old man) but I will get either Rory or Eric to lab this tomorrow.

[LRL]

No worries. Sheesh, sounds like a bad week for everyone.

You guys take it easy it's not the end of the world.

[sirhc]

No worries. Sheesh, sounds like a bad week for everyone.

You guys take it easy it's not the end of the world.

I appreciate your understanding buddy, I am really sorry and I will get it done tomorrow I hope!

[Eric Stern]

To summarize the problem, bridges after the switch don't receive BPDU updates correctly(or at all) if stp is not enabled in the switch.

Prior to 1.1.0rc19 the BPDU packets were being forwarded by the switch and copied to the CPU, this was causing the BPDU "leaks". In rc19 we changed it so that BPDUs are redirected to the CPU and not forwarded. So then if STP is disabled the BPDUs basically get swallowed by the switch. After doing some research it seems that different switches do different things in this case, some swallow the packets, some forward them. No one seems to know what the "right" thing to do is.

Anyways, I've changed this in rc21 (Chris will be sending this to you shortly). Both STP and LACP use BPDUs, so now if both STP and LACP are disabled then the switch will forward BPDUs, otherwise it will swallow them. This should fix your problem.

I CAN NOT RELEASE RC21 AS IT WILL CAUSE BPDU PACKET BLEED ACROSS VLANs! - CHRIS

[sirhc]

From what I have read managed switches with VLANs should not pass BPDU packets unless they have MSTP enabled as this would cause BPDU packet bleed across VLANs causing very flaky and wacky behavior which would mean that we are handling it correctly but I agree that a DUMB switch will pass BPDU packets but it does not have VLANs in the equation.

But to be clear there is not a BUG with v1.1.0rc19 with the way it is handling BPDU packets currently.

So far v1.1.0rc19 has proven very stable with what feedback we have received so far.

To be able to do what you're asking LRL we need to enable MSTP which is on the road map but not slated for v1.1.0 but SOON.

I will ask Eric to look into enabling MSTP in the near future as the switch is capable of this feature.

Please call me if you would like to talk about this but if we revert to the way BPDU packets were handled previously it would cause BPDU packets to cross over VLANs which would break things for more people then it would help.

Now we could allow BPDU packets to pass through the switch like a DUMB switch if the following criteria is met:
NO VLANs are configured
R/STP is DISABLED
NO LACP or STATIC LAGS are configured
LOOP PROTECTION is DISABLED

But honestly I think you really want MSTP which is COMING SOON

[sirhc]

If anyone else would like to weigh in on this debate please feel free to comment.

[LRL]

From what I have read managed switches with VLANs should not pass BPDU packets unless they have MSTP enabled as this would cause BPDU packet bleed across VLANs causing very flaky and wacky behavior which would mean that we are handling it correctly but I agree that a DUMB switch will pass BPDU packets but it does not have VLANs in the equation.

But to be clear there is not a BUG with v1.1.0rc19 with the way it is handling BPDU packets currently.

So far v1.1.0rc19 has proven very stable with what feedback we have received so far.

To be able to do what you're asking LRL we need to enable MSTP which is on the road map but not slated for v1.1.0 but SOON.

I will ask Eric to look into enabling MSTP in the near future as the switch is capable of this feature.

Please call me if you would like to talk about this but if we revert to the way BPDU packets were handled previously it would cause BPDU packets to cross over VLANs which would break things for more people then it would help.

Now we could allow BPDU packets to pass through the switch like a DUMB switch if the following criteria is met:
NO VLANs are configured
R/STP is DISABLED
NO LACP or STATIC LAGS are configured
LOOP PROTECTION is DISABLED

But honestly I think you really want MSTP which is COMING SOON

It would probably be best to pass BPDUs along if no other protocols are enabled that use them. This prevents the switch when being dumb, from causing loops that STP on other devices would normally resolve.

I'm certainly not fully versed on Loop Protection, but does it use BPDUs?

[LRL]

If I downgrade back to 1.0.10 from1.1.0rc19 is there any expected issues?

[josh]

There is a bug in Chrome/Firefox with selecting POE on Port 1. I'm trying to replicate it. I can make it go away by refreshing the page. It's frickin annoying

[adairw]

There is a bug in Chrome/Firefox with selecting POE on Port 1. I'm trying to replicate it. I can make it go away by refreshing the page. It's frickin annoying

What exactly is the bug? I just logged in to two switches and they both do the same thing. They are 48VH right now, and I can't select anything but off. Same thing for ports 2, 3 & 4 though.
If the ports are off already I can of course select whatever.