LAG not working properly anymore

[Accel Wireless]

Two days ago it quit working and nothing was changed. The counters on the LAG tab for these two ports quit incrementing. Everything is set for LACP-A like all other towersites and exact configurations. We have rebooted the mikortik and it still will not show load balancing across the two LACP ports. IT just shows traffic on one or the other port but not both. if we shut down a port on the mikrotik it forces traffic to the other port and visa versa. The only thing we have not tried is rebooting the Netonix switch to see if that clears the issue but we wanted to check here for any other cause or something we need to look at.


Software Ver 1.4.7

[sirhc]

First upgrade to v1.4.9

Second notice the settings at the top of LACP

If you select source or destination MAC you might get all traffic on one port if all the traffic is routed traffic because the source and destination MAC would remain the same.

You only use MAC the defining balance on a Layer 2 FLAT SEGMENT.

If you have a router hooked to the switch and the traffic is routed then you need to use source and or destination IP as the criteria to load balance.

[Accel Wireless]

But the fact it has been working fine until 2 days ago and is still working on 12 other sites exactly the same way is baffling? Did you look at the attachment on our current setting?

[sirhc]

I did not see that attachment for whatever reason the first time?

Try a reboot like you suggested but not sure that will do anything.

Also show me a screen grab of Status TAB with a graph for each port in LAG

[sirhc]

By the way if the majority of the traffic is coming from the Router to the Switch than the problem is in the router.

Reason why is LACP between the MT Router and the Netonix switch divvies up the traffic as follows:
Traffic from Router to Switch is divvied up across the 2 links on the router side
Traffic from the Switch to the Router is divvied up across the 2 links on the switch side.

LACP on each device controls the OUTFLOW or EGRESS of traffic on that device to the other device.

[Accel Wireless]

We rebooted the switch and the problem is gone now.. Both LAG ports on the Mikrotik now show traffic and the Netnoix counters are incriminating as normal. We also found another switch that would give us an error when trying to save and commit a change so we rebooted that one as well and the problem was corrected so now we can save changes again.

[Accel Wireless]

Here is what I think you were asking for. I didn't see this post until after I rebooted the unit in question..

[sirhc]

Next time you run into the save issue try another computer, sometimes corrupted cached files in your browser can cause this.

Also next time if this LACP LAG issues occurs try rebooting the router and see if that also fixes the issues. LACP is a negotiated protocol between both devices so when you reboot the switch the router also is forced to renegotiate the LAG. I am not saying this was not a switch issue but I have never seen this and I run LAGs on all my switches. Issue could have been Netonix or MT.

If the majority of the traffic from the router to the switch was shifted to 1 port but the traffic from the switch to the router was being split across the 2 links then the issue would point to the MT router since each device on each side of the LAG only divvies up the egress traffic leaving that device to the other device.

If under the LAG TAB both ports showed active for the LAG then the switch should have been divvying up the traffic going to the router across both ports. Keep in mind the traffic is never symmetrical as it does so on a per stream not per packet basis.

[Accel Wireless]

I tried all of the above you mentioned, rebooting the router was the first thing I tried as well as double checking and comparing settings to our other 40 sites. All settings were exactly the same The only thing that work was to reboot the netonix via the GUI.

zero traffic was going over the other link it was all shifted to the one link or the other but never both until after I rebooted the Netonix. Its almost as if the Netonix did not know the other link was available or part of LAG?

Hopefully it never happens again but if it does I will know what to do and maybe you would like to log in and look at it?

[sirhc]

If it happens again and the firmware is up to date I could look at it yes.