Double throughput with Bridge, Lagg or other?
-
"Only unassigned physical ports can be added to a LAGG" so you have to remove it from LAN first. The firewall rules will remain attached to LAN. After all this you'll have LAN assigned to the LAGG interface. Step 2 in that doc is adding the second interface to the LAGG. Step 7 is adding the interface that was "in" LAN to LAGG. Just be careful in all this you don't disconnect yourself. It's been a while, but I've done this in a remote data center however was using an opt interface.
-
Yup, you have to unassign those interfaces so you will want to be connected to the firewall via the WAN or a management interface when you try that.
Steve
-
So the document is misleading me when it says "Now both re0 and re2 are members of a LAGG and that LAGG is the LAN interface with all of the existing configuration in place"? I'm still not getting how any existing configurations can remain in place if all the interfaces added have to be unassigned.
Question #1; Won't unassigning an interface remove all it's configurations?
Question #2; Why is it letting me add PHYSICAL interfaces igb0 & igb1 that ARE assigned?
These are the two issues I'm trying to get answered. I know unassigning them will work, that's obvious. But then I'll be left with hours of reconfiguring the unassigned interface. I'm trying to avoid many hours of work.
-
Create the lagg with just one unassigned NIC.
Re-assign LAN from the physical NIC to the LAGG. LAN config remains.
Add the, now unassigned, NIC that was LAN to the LAGG.
Now you have both NICs in the LAGG and the LAGG assigned as LAN with the same config it had before.
Steve
-
@stephenw10 Don't know why I got hung up on your step #2 which is the doc's step #3. Seems obvious now. Thought it meant something else so wasn't looking where I shoulda been. Uggg! Working now, thanks to you and everyone.
-
After running igb2 and igb3 as dynamic active LACP members I find that the GUI Lan "Traffic Graph" shows 2G max, however the Status, Monitoring for the Lan adapter only shows 1G max. And an SNMP probe, via PRTG, only shows 1G max. Is this by design or a shortcoming?
-
Unfortunately there have been a number of doubling or halving bugs affecting the graph data over the years. And it's usually the traffic graphs that are wrong. For example:
https://redmine.pfsense.org/issues/10812
Though I think that only affected 2.5. And it's now resolved.Are you actually seeing ~2Gbps if you test from something other than the firewall?
How are you actually generating that traffic?
Steve
-
I'm running 2.4.5, haven't applied 2.4.5_1 yet. I'm looking at actual customer traffic. LACP was applied because the single adapter Lan was maxing out during peak hour load. With LACP now active I'm seeing 1G - 1.3G of actual traffic. Whatever the GUI, or the sum of both switch ports reports, the GUI monitor page or SNMP only sees 1/2 of this value. Was there a fix with _1?
-
Nothing listed: https://docs.netgate.com/pfsense/en/latest/releases/2-4-5-p1.html
But there might be other things not shown there.
You should update to 2.4.5p1 anyway for all the other things that are fixed there.
Steve
-
Upgrading to 2.4.5_1 appears to have corrected the LACP 1/2 throughput issue.
-
Ah, good to hear. Thanks for reporting back.
