CakeQOS CakeQoS-Merlin v2.0.0 alpha

[MvW]

spdMerlin.log

Apr 14 13:40:56 RT-AC86U spdMerlin: Setting QoS Upload Speed to 18106 Kbps (was 19418 Kbps)
Apr 14 14:10:01 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for WAN interface
Apr 14 14:10:29 RT-AC86U spdMerlin: Speedtest results - Download: 131.77 Mbps (data used: 118.2 MB) - Upload: 15.52 Mbps (data used: 14.4 MB)
Apr 14 14:10:29 RT-AC86U spdMerlin: Connection quality - Latency: 9.22 ms (0.89 ms jitter) - Packet Loss: 0.0%
Apr 14 14:10:29 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for VPNC1 interface
Apr 14 14:11:01 RT-AC86U spdMerlin: Speedtest results - Download: 65.31 Mbps (data used: 98.1 MB) - Upload: 11.15 Mbps (data used: 18.0 MB)
Apr 14 14:11:01 RT-AC86U spdMerlin: Connection quality - Latency: 10.38 ms (0.31 ms jitter) - Packet Loss: 0.0%
Apr 14 14:11:06 RT-AC86U spdMerlin: Calculated Download speed (137297) Kbps does not exceed 5% threshold of existing value (143608 Kbps)
Apr 14 14:11:06 RT-AC86U spdMerlin: Calculated Upload speed (17741) Kbps does not exceed 5% threshold of existing value (18106 Kbps)
Apr 14 14:40:01 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for WAN interface
Apr 14 14:40:27 RT-AC86U spdMerlin: Speedtest results - Download: 133.67 Mbps (data used: 125.8 MB) - Upload: 15.88 Mbps (data used: 14.6 MB)
Apr 14 14:40:27 RT-AC86U spdMerlin: Connection quality - Latency: 5.66 ms (1.15 ms jitter) - Packet Loss: Not available.
Apr 14 14:40:27 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for VPNC1 interface
Apr 14 14:40:51 RT-AC86U spdMerlin: Speedtest results - Download: 113.32 Mbps (data used: 139.8 MB) - Upload: 14.48 Mbps (data used: 12.9 MB)
Apr 14 14:40:51 RT-AC86U spdMerlin: Connection quality - Latency: 9.59 ms (4.34 ms jitter) - Packet Loss: 0.0%
Apr 14 14:40:56 RT-AC86U spdMerlin: Setting QoS Download Speed to 136311 Kbps (was 143608 Kbps)
Apr 14 14:40:56 RT-AC86U spdMerlin: Calculated Upload speed (17336) Kbps does not exceed 5% threshold of existing value (18106 Kbps)
Apr 14 15:10:01 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for WAN interface
Apr 14 15:10:27 RT-AC86U spdMerlin: Speedtest results - Download: 124.16 Mbps (data used: 135.6 MB) - Upload: 16.59 Mbps (data used: 9.8 MB)
Apr 14 15:10:27 RT-AC86U spdMerlin: Connection quality - Latency: 9.33 ms (6.03 ms jitter) - Packet Loss: 0.0%
Apr 14 15:10:27 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for VPNC1 interface
Apr 14 15:11:00 RT-AC86U spdMerlin: Speedtest results - Download: 101.16 Mbps (data used: 153.6 MB) - Upload: 11.05 Mbps (data used: 19.0 MB)
Apr 14 15:11:00 RT-AC86U spdMerlin: Connection quality - Latency: 11.51 ms (7.57 ms jitter) - Packet Loss: 0.0%
Apr 14 15:11:05 RT-AC86U spdMerlin: Calculated Download speed (134279) Kbps does not exceed 5% threshold of existing value (136311 Kbps)
Apr 14 15:11:05 RT-AC86U spdMerlin: Setting QoS Upload Speed to 17143 Kbps (was 18106 Kbps)
Apr 14 15:40:02 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for WAN interface
Apr 14 15:40:30 RT-AC86U spdMerlin: Speedtest results - Download: 263.09 Mbps (data used: 392.1 MB) - Upload: 28.75 Mbps (data used: 14.5 MB)
Apr 14 15:40:30 RT-AC86U spdMerlin: Connection quality - Latency: 8.38 ms (0.78 ms jitter) - Packet Loss: 0.0%
Apr 14 15:40:30 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for VPNC1 interface
Apr 14 15:40:57 RT-AC86U spdMerlin: Speedtest results - Download: 120.82 Mbps (data used: 201.7 MB) - Upload: 26.64 Mbps (data used: 24.4 MB)
Apr 14 15:40:57 RT-AC86U spdMerlin: Connection quality - Latency: 10.45 ms (5.83 ms jitter) - Packet Loss: 0.0%
Apr 14 15:41:01 RT-AC86U spdMerlin: Setting QoS Download Speed to 146118 Kbps (was 307200 Kbps)
Apr 14 15:41:01 RT-AC86U spdMerlin: Setting QoS Upload Speed to 18294 Kbps (was 30720 Kbps)
Apr 14 16:10:02 RT-AC86U spdMerlin: Starting speedtest using using auto-selected server for WAN interface

As far as I can see spdMerlin is already done at the time you mentioned.

When the stats don't show up, if I hit Apply on the CakeQos-Merlin tab, without making any changes, the stats show up after the router is done processing the (none) changes.

Are there any other processes of interest? Otherwise I'm leaning towards uninstalling uiScribe/scribe/syslogn-ng, so we can see things in their context.

 

[MvW]

I made a separate post in the spdMerlin thread, as I just discovered another issue, and I'm not sure whether related:

spdMerlin - spdMerlin - v4.x - speedtest and bandwidth monitoring

So is it possible that for every Speedtest time it automatically do multiple Speedtests concurrently? Wouldn't 3 speedtests concurrently affect your maximum achievable speed? I assume each speedtest tries to utilize the bandwidth available as optimal as possible, hence running 3 speedtests at...

www.snbforums.com

 

[XIII]

My ISP increased my download/upload speed (from 250/25 to 300/30), but I could not achieve those, until I realized that Cake was active with limits set to 95% of 250/25...

Unfortunately (95% of) 300/30 seems to much to handle with Cake for my AC86u, so my days with Cake might be over?

 

[MvW]

My ISP increased my download/upload speed (from 250/25 to 300/30), but I could not achieve those, until I realized that Cake was active with limits set to 95% of 250/25...

Unfortunately (95% of) 300/30 seems to much to handle with Cake for my AC86u, so my days with Cake might be over?

I think we might have the same ISP...

 

[dave14305]

As far as I can see spdMerlin is already done at the time you mentioned. When the stats don't show up, if I hit Apply on the CakeQos-Merlin tab, without making any changes, the stats show up after the router is done processing the (none) changes. Are there any other processes of interest? Otherwise I'm leaning towards uninstalling uiScribe/scribe/syslogn-ng, so we can see things in their context.

It's like cake isn't being started properly after the test, but there's no evidence of this other than the stats collection doesn't find any cake qdiscs. Hitting Apply forces a qos restart, so it hides the root cause of the problem.

I would speculate that the AutoBW qos restart at 15:11 in your logs didn't properly run /tmp/qos start or it had an error this time. The next time you see this happen, check the content of cat /etc/cake-qos.conf

I pushed an update to the alpha branch that will give an error message if Cake isn't active for both upload and download. Once we better understand what causes that scenario, we might be able to handle it better.

 

[Wade Coxon]

Not sure about the rest of you, but when I last tested, my rt-ac86u would reboot any time I set QoS from a disabled state to an enabled state with cake. The change took effect, but caused the reboot to occur.

Perhaps the automate speed tests are experiencing a similar difficulty with re-enabling QoS, just without the obvious reboot occurring, and instead failing to re-enable QoS.

 

[heysoundude]

My ISP increased my download/upload speed (from 250/25 to 300/30), but I could not achieve those, until I realized that Cake was active with limits set to 95% of 250/25...

Unfortunately (95% of) 300/30 seems to much to handle with Cake for my AC86u, so my days with Cake might be over?

have you tried closer to 90% of the higher speeds? ;-D

 

[XIII]

have you tried closer to 90% of the higher speeds? ;-D

Yes.

 

[TonyK132]

If you are running 384.19 you have to use 1.0.7

OK. Used the install script from the Wiki and that got me back to 1.0.7 after a couple of reboots. Seems to work the same as 1.0.8.

 

[heysoundude]

Yes.

OK. where are your settings at. other than speeds?

 

[XIII]

OK. where are your settings at. other than speeds?

All just (Asuswrt-Merlin) defaults:

• WAN packet overhead = 18
• MPU = 64
• Mode = Normal

 

[docious]

Can anyone confirm if the RT-AC67U is the Asus HND platform, necessary for the cake-qos alpha branch? I've had a decent online rummage but it isn't clear. Thanks.

CPU Model	BCM470x - Cortex A7 ARMv7 revision 0 - Rev. c0 (Cores: 2)

 

[MvW]

Can anyone confirm if the RT-AC67U is the Asus HND platform, necessary for the cake-qos alpha branch? I've had a decent online rummage but it isn't clear. Thanks.

​

CPU Model	BCM470x - Cortex A7 ARMv7 revision 0 - Rev. c0 (Cores: 2)

Unfortunately, as the RT-AC68U is not, yours is not either, as far as I can tell.

Upgrade to the ac66u??

Not sure if it was confirmed by manufacturer but I did see some images. To a AC67U and wanted to know I the images are true. I reason I ask is the new 67u does not have external antennae and that is 1 feature I would like to have on my next wireless router along with USB 3.0. Any thoughts or...

www.snbforums.com

 

[Zavar]

I've been using CakeQOS since updating to the 386.2 Merlin release and it has been working quit well, or I haven't really noticed any issues. I just started taking a closer look at the stats and I've noticed quite a few drops for both the download and uploads. What is a "good" packet to drop ratio (is this the best way to look at it?)? I'm assuming to get better results, I should further tune my download/upload bandwidth settings, or are there other settings that I should also be looking at?

The other thing that I noticed, in the stats sections are the threshold rates, in my setup these show being above the bandwidth settings that I've used. Should I be adjusting decreasing the bandwidth settings so that threshold rates are actually at my available tested upload/download rates?

Thanks for any help!

 

[dave14305]

The other thing that I noticed, in the stats sections are the threshold rates, in my setup these show being above the bandwidth settings that I've used. Should I be adjusting decreasing the bandwidth settings so that threshold rates are actually at my available tested upload/download rates?

Please post a screenshot of your configured rates and the threshold stats. They should not be higher.

 

[Zavar]

I hope this is what you need...

 

[dave14305]

I hope this is what you need...

View attachment 33365
View attachment 33366
View attachment 33367

You’ve stumbled upon one of the nerd controversies with Asus firmware. The QoS page takes the bandwidth input labeled Mb/s (Megabits per second), but stores it in nvram as kilobits as a multiple of 1024 instead of 1000. Google “mebibit” versus “megabit”. The tc command that underlies all these QoS choices is given the nvram values as kilobits (x 1000) and reports it as such.

This was recently discussed and the feeling was it was better to leave it as-is for consistency with the rest of the firmware units of measurement.

 

[Zavar]

Ah OK, I was wondering if it was a "unit" discrepancy. Thank You!

Any comment re. dropped stats?

 

[Jack Yaz]

You’ve stumbled upon one of the nerd controversies with Asus firmware. The QoS page takes the bandwidth input labeled Mb/s (Megabits per second), but stores it in nvram as kilobits as a multiple of 1024 instead of 1000. Google “mebibit” versus “megabit”. The tc command that underlies all these QoS choices is given the nvram values as kilobits (x 1000) and reports it as such.

This was recently discussed and the feeling was it was better to leave it as-is for consistency with the rest of the firmware units of measurement.

What's the saying about waiting ages for a bus and then 2 come at once?

 

[randomName]

Would also like to see more discussion about drop rates. At one point before a router refresh I had 300k dl/drops, and 150k upl/drops. Tons of ACK drops also