Can you run a speed test or two now and send a screenshot of netdata graphs?
I saved it as PDF, see if you can access it...
https://drive.google.com/file/d/1MUoSOvpEZDrS2AW-PoO0dOgoSlFq6g5d/view?usp=sharing
I was able to access and review it. Couple questions about your connection… are you by any chance using a cable modem with the infamous Puma 6 chipset?
Yes WOW cable internet using an Arris Touchstone(motorola) CM8200
my bufferbloat when I set my download to 150mbps or less
as soon as I raise that even by 50, back to the 100 plus milliseconds when using cake.
At this point maybe something with my equipment, provider, haunted house or not sure. Odd thing is FQ Codel/ simple qos gives me the results below...seem to be working better than cake form me at 90% bandwith...
so I guess ill rock with that for now
I am game if you guys want to tinker a bit more...let me know ill test stuff if you want me to, but i don't want to waste your guys time. I am sure you guys have more important things to do
@moeller0 Could this be something with needing to disable GRO/GSO?
Cake by default dissects GSO/GRO aggregates into their constituting segments for shaper rates < 1Gbps, which essentially should deal with that issue, at the cost of slightly more CPU usage....
I’m grasping because this behavior is perplexing to me. I have to be missing something obvious.
So if this is a puma chipset I think this probably explains your problem the modem itself is incapable of handling the speeds without delays
Looked up, not puma.
root@Willthetech_Home:~# cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3 CPU4 CPU5 CPU6 CPU7
0: 28 0 0 0 0 0 0 0 IO-APIC 2-edge timer
1: 0 8 0 0 0 0 0 0 IO-APIC 1-edge i8042
4: 0 0 0 15 0 0 0 0 IO-APIC 4-edge ttyS0
8: 0 0 1 0 0 0 0 0 IO-APIC 8-edge rtc0
9: 0 5 0 0 0 0 0 0 IO-APIC 9-fasteoi acpi
12: 9 0 0 0 0 0 0 0 IO-APIC 12-edge i8042
123: 0 0 0 0 0 47 0 0 PCI-MSI 32768-edge i915
124: 0 49 0 0 0 0 2092 0 PCI-MSI 376832-edge ahci[0000:00:17.0]
125: 0 0 0 0 0 0 0 0 PCI-MSI 327680-edge xhci_hcd
126: 1 0 0 0 0 0 0 0 PCI-MSI 2097152-edge eth0
127: 0 616 0 0 881166 0 0 0 PCI-MSI 2097153-edge eth0-TxRx-0
128: 0 0 26 0 0 0 0 739590 PCI-MSI 2097154-edge eth0-TxRx-1
129: 0 0 0 563529 0 0 0 0 PCI-MSI 2097155-edge eth0-TxRx-2
130: 0 633689 0 0 13 0 0 0 PCI-MSI 2097156-edge eth0-TxRx-3
131: 0 0 0 0 0 648187 0 0 PCI-MSI 2097157-edge eth0-TxRx-4
132: 0 1978723 0 0 0 0 13 0 PCI-MSI 2097158-edge eth0-TxRx-5
133: 0 0 0 0 0 0 773355 16 PCI-MSI 2097159-edge eth0-TxRx-6
134: 15 0 709451 0 0 0 0 0 PCI-MSI 2097160-edge eth0-TxRx-7
135: 0 0 0 0 0 0 0 0 PCI-MSI 2099200-edge eth1
136: 0 2854 0 0 0 0 7 0 PCI-MSI 2099201-edge eth1-TxRx-0
137: 0 0 2854 0 0 0 0 7 PCI-MSI 2099202-edge eth1-TxRx-1
138: 7 0 0 0 0 0 0 2854 PCI-MSI 2099203-edge eth1-TxRx-2
139: 2854 7 0 0 0 0 0 0 PCI-MSI 2099204-edge eth1-TxRx-3
140: 0 0 7 0 0 0 0 2854 PCI-MSI 2099205-edge eth1-TxRx-4
141: 0 0 0 7 0 2854 0 0 PCI-MSI 2099206-edge eth1-TxRx-5
142: 0 0 0 2854 7 0 0 0 PCI-MSI 2099207-edge eth1-TxRx-6
143: 0 0 0 0 0 7 2854 0 PCI-MSI 2099208-edge eth1-TxRx-7
144: 0 0 0 0 0 0 0 0 PCI-MSI 2101248-edge eth2
145: 0 0 0 0 0 2854 0 7 PCI-MSI 2101249-edge eth2-TxRx-0
146: 7 0 2854 0 0 0 0 0 PCI-MSI 2101250-edge eth2-TxRx-1
147: 0 7 0 0 0 0 2854 0 PCI-MSI 2101251-edge eth2-TxRx-2
148: 2854 0 7 0 0 0 0 0 PCI-MSI 2101252-edge eth2-TxRx-3
149: 0 0 0 7 2854 0 0 0 PCI-MSI 2101253-edge eth2-TxRx-4
150: 0 0 0 2854 7 0 0 0 PCI-MSI 2101254-edge eth2-TxRx-5
151: 0 0 0 0 0 7 0 2854 PCI-MSI 2101255-edge eth2-TxRx-6
152: 0 0 2854 0 0 0 7 0 PCI-MSI 2101256-edge eth2-TxRx-7
153: 0 0 0 0 0 0 0 1 PCI-MSI 2103296-edge eth3
154: 90 0 0 0 1038997 0 0 0 PCI-MSI 2103297-edge eth3-TxRx-0
155: 0 128 0 946438 0 0 0 0 PCI-MSI 2103298-edge eth3-TxRx-1
156: 0 0 64 0 0 0 0 794616 PCI-MSI 2103299-edge eth3-TxRx-2
157: 0 622143 0 142 0 0 0 0 PCI-MSI 2103300-edge eth3-TxRx-3
158: 0 0 0 0 132 1021147 0 0 PCI-MSI 2103301-edge eth3-TxRx-4
159: 0 0 0 0 0 578063 0 0 PCI-MSI 2103302-edge eth3-TxRx-5
160: 0 0 0 0 0 0 944234 0 PCI-MSI 2103303-edge eth3-TxRx-6
161: 0 0 0 0 0 0 1012058 110 PCI-MSI 2103304-edge eth3-TxRx-7
NMI: 0 0 0 0 0 0 0 0 Non-maskable interrupts
LOC: 1953195 1995905 1706428 1551161 1809084 1741130 2127408 1639369 Local timer interrupts
SPU: 0 0 0 0 0 0 0 0 Spurious interrupts
PMI: 0 0 0 0 0 0 0 0 Performance monitoring interrupts
IWI: 0 0 0 0 0 0 0 0 IRQ work interrupts
RTR: 4 0 0 0 0 0 0 0 APIC ICR read retries
RES: 1767 2392 2446 1893 1924 3442 1434 1674 Rescheduling interrupts
CAL: 1179756 933645 667503 556902 1049778 765585 1303980 812258 Function call interrupts
TLB: 89 172 183 96 214 148 130 152 TLB shootdowns
TRM: 0 0 0 0 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 0 0 0 0 Threshold APIC interrupts
DFR: 0 0 0 0 0 0 0 0 Deferred Error APIC interrupts
MCE: 0 0 0 0 0 0 0 0 Machine check exceptions
MCP: 19 20 20 20 20 20 20 20 Machine check polls
ERR: 4
MIS: 0
PIN: 0 0 0 0 0 0 0 0 Posted-interrupt notification event
NPI: 0 0 0 0 0 0 0 0 Nested posted-interrupt event
PIW: 0 0 0 0 0 0 0 0 Posted-interrupt wakeup event
root@Willthetech_Home:~#
I think my next steps here are what @elan suggested. Check the wiring and probably the modem. I am going to run new coaxial from the outside to the modem inside and keep it short, right now is kind of long, probably over 12 feet...maybe noise is making this an issue. I am going to try to shorten it to less than 6 feet then run new cat6a from the modem to the openwrt box. then test it there... I'm still confused why FQ_Codel / simple QOS is giving me low 50's.We will see ill update here gain with my results. I really appreciate all of your help/suggestions/tinkering. Happy new year guys
OK, so yes your system does allow frequency scaling, but it is not clear that during the speedtest it actually scales back at all....
The interrupt counts look okay-ish, no huge imbalance between cores. It seems your NICs are all 8-queue devices, but that really should not matter... Then again, for a qdisc that is restricted to a single CPU like cake distributing the 8 queues per NIC over all CPUs might not be ideal (but should not matter too much).
For coaxial cable in a normal cable plant an additional 12 feet should not matter, unless they are of bad quality and/or the connectors are bad.
Yes, that is what is puzzling here
@WilR You could check your cable modem status page and see if you have an abnormally high number of uncorrectables (uncorrectable code words). Might also want to check your signal levels are in range. That shouldn’t be the immediate cause of the issues we are seeing here, but would speak to whether there might be other ISP last mile issues.
One idea would be to remove the cable modem entirely. Just put a "server" on the "wan" side (directly plugged into the ethernet port), and a "client" on the LAN side and then run iperf3 between the two of them.
Of course, this requires that you have at least 2 computers you can set up iperf3 on to be the testing machines. But if you do this and are able to shape up to a gigabit, then you'll know the issue is in your ISP not the cpu usage or NICs or anything like that.