side note -
(one of the kids got a new job which requires a lot of video meetings which are interviews, and Q&A of people a few days a week..) So I wanted to do something that she would be (insert emotion instead of logical description here..)
I had and mt3000 and an external unmanaged netgear 5 port and we're a 300/30 with an SB8200 (docsis 3.1) .. I talked myself into an mt6000 and said it was b/c of the integrated switch ports (understanding there is no switch chip..)
I put my eero 6+ APs (in bridge mode) into the switch ports of the mt6000 (latest snapshot with apk) and moved on with life.. A couple of the meetings she got a 'low bandwidth' message.. so I setup openspeed test on a machine at home.. and wasn't getting 150Mbs up/down internally.. being confused.. I put the APs back into the netgear.. and was getting 500's.. putting the APs back into the mt6000.. back to 150's..
at the end of this.. I see all 16 of the fq_codel mq's (1-f) and when I look at them (tc -s qd show dev eth0)
qdisc fq_codel 0: parent :7 limit 10240p flows 1024 quantum 1518 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 430572026546 bytes 361326769 pkt (dropped 1225, overlimits 0 requeues 54407)
backlog 0b 0p requeues 54407
maxpacket 24288 drop_overlimit 1216 new_flow_count 4285992 ecn_mark 0 drop_overmemory 1216
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: parent :6 limit 10240p flows 1024 quantum 1518 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 1104423270 bytes 813526 pkt (dropped 0, overlimits 0 requeues 53)
backlog 0b 0p requeues 53
maxpacket 1518 drop_overlimit 0 new_flow_count 20927 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: parent :5 limit 10240p flows 1024 quantum 1518 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 129108598 bytes 97933 pkt (dropped 0, overlimits 0 requeues 10)
backlog 0b 0p requeues 10
maxpacket 1518 drop_overlimit 0 new_flow_count 487 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: parent :4 limit 10240p flows 1024 quantum 1518 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 122151263 bytes 86256 pkt (dropped 0, overlimits 0 requeues 9)
backlog 0b 0p requeues 9
all the other child queues are blank..
so I put a queue on br-lan..
qdisc fq_codel 800e: root refcnt 2 limit 20480p flows 2048 quantum 1514 target 10ms interval 100ms memory_limit 64Mb drop_batch 1
Sent 7750947713 bytes 5832598 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
maxpacket 4542 drop_overlimit 0 new_flow_count 42688 ecn_mark 0
new_flows_len 0 old_flows_len 0
My questions are:
of the 16 eth0 queue's.. why are most of them not being used.. I'm also confused how they were intended to be used..
I don't think under any circumstances I should have dropped packets at the ethernet port from what we're doing here.. and I cannot find any definitive definition of what the requeue is for..
I'm causing those big spikes on purpose pulling/pushing things from work.. which has more bw than home..
I'm running qosmate on eth1 and things seem fine there.. (tm)
Also burned more money on an SG108E because it said it works with dscp tagged packets.. and see no errors on the interfaces (latest firmware as well)
Is there something else that I could/should be doing..
I've not seen anyone else talking about this.. so I'm not sure if I found something.. or broke something..
One and three are interviews.. two is the boy playing a game/discord then going to work.. unmarked big spike on the left is the boy on facetime with his friends..
She is using a macbook air.. the eero 6+ are mesh/standalone except for one..
(of all the house projects there are to do.. I can't get time to setup back hauls on the 6+'s.. yet)
cpu graph for reference..
(too much information? not enough?)
Thanks in advance.