v1.4.8 Bug reports and Comments

[sirhc]

Nothing changed between v1.4.7 and v1.4.8 to cause this issue.

Sure you can downgrade but I would not as this is NOT the issue.

You provide little information to assist you.

You have a lot of these switches, what model is this we are talking about?

What ports are in the LAG to the router, if they are SFP ports I would look at SFP modules.

Maybe put up some information to help you suchs as the following TABs:
Status
Ports
STP
LAG
VLANs

Put up a portion of the Log showing this event.

[lweidig]

WS-26-400-AC switch connecting to a CCR1009, here are the tabs for the ports with the issue:

Log file - 100's / 1000's of these:

Dec 11 21:07:51 LACP: LACP changed state to Active on port 1 (key 100)
Dec 11 21:07:51 switch[2866]: LACP changed state to Active on port 1 () (key 100)
Dec 11 21:07:53 sSMTP[2871]: Sent mail for xxx@excel.net (221 Service closing transmission channel) uid=0 username=xxx outbytes=574

The ports are connected with 3' CAT6 cables in a rack.

[sirhc]

So these are the only things I can think to try:


I know you say the cables are good cables but try differnt cables. CAT6 is not needed, use simple CAT5e. The reason I say this is some CAT6 Cables put stress on the RJ45 connectors being thicker heavier cable. - GRASPING AT STRAWS but I have seen bad cables do this.

Make sure there is nothing wrong with one of the ports in one of the switches, try differnt ports in the LACP LAG one at a time.

If you have a spare switch swap one out at a time.

If you try all the above and no love I would look at the router.

Your setup looks fine so it has to be something in cabling, a bad port, a bad switch, or a bad port in the router or something with the router.

You could also use a Static LAG just to see if the issue goes away which may point to how the router is doing LACP? I know we are handling LACP right as this was tested in our LABs with other switches. I also run LACP all over my net at my WISP with no issue.

[lweidig]

Was able to perform more diagnostics on this over the evening. Cable swaps has no affect on the situation. During this though we noticed one of the ports (not part of the LACP group) had much higher response times than the others. It was connected to a Cambium ePMP 2000. When that device was power cycled ALL of the issues went away!

Now, why would this happen? Seems there is some sort of issue that a single port could affect other ports so significantly.

If possible Chris can you move these posts off to their own topic so the discussion can continue with it not being part of the 1.4.8 issues topic. Thanks!

[sirhc]

Was able to perform more diagnostics on this over the evening. Cable swaps has no affect on the situation. During this though we noticed one of the ports (not part of the LACP group) had much higher response times than the others. It was connected to a Cambium ePMP 2000. When that device was power cycled ALL of the issues went away!

Now, why would this happen? Seems there is some sort of issue that a single port could affect other ports so significantly.

I am thinking this has something to do with STP?

You should disable STP on ports facing AP's, and make sure STP is disabled in the radio itself:
viewtopic.php?f=6&t=2748&p=19018&hilit=+RSTP+disable#p19018
viewtopic.php?f=17&t=1480&p=16549&hilit=+RSTP+disable#p16549
viewtopic.php?f=17&t=1480&p=11011&hilit=+RSTP+disable#p11011

[lweidig]

Had actually tries turning STP off completely on the switch and was still experiencing this. I will review the topics though to see if that should be a best practice going forward as we use these at ALL sites.

[sirhc]

Had actually tries turning STP off completely on the switch and was still experiencing this. I will review the topics though to see if that should be a best practice going forward as we use these at ALL sites.

So we are back to these suggestion to try.

So these are the only things I can think to try:
I know you say the cables are good cables but try differnt cables. CAT6 is not needed, use simple CAT5e. The reason I say this is some CAT6 Cables put stress on the RJ45 connectors being thicker heavier cable. - GRASPING AT STRAWS but I have seen bad cables do this. - TRIED THIS NO HELP

Make sure there is nothing wrong with one of the ports in one of the switches, try differnt ports in the LACP LAG one at a time.

If you have a spare switch swap one out at a time.

If you try all the above and no love I would look at the router.

Your setup looks fine so it has to be something in cabling, a bad port, a bad switch, or a bad port in the router or something with the router.

You could also use a Static LAG just to see if the issue goes away which may point to how the router is doing LACP? I know we are handling LACP right as this was tested in our LABs with other switches. I also run LACP all over my net at my WISP with no issue.