CakeQOS CakeQoS-Merlin v2.0.0

[Morris]

The first thing I'd try is besteffort in both directions. You have no controlee of what goes in what TIN and this will avoid other traffic having priority over the VPN. You must also consider what traffic is in the VPN as a download could cause the entire VPN to be slowed for other traffic such as net controls. VPN might not be a good fit for cake unless it has it's own cake interface and that's not implemented.

Good luck,
Morris

Second thought, you stated all traffic is tunneled and no split tunneling allowed. QOS on the router can not help you. If you VPN has the feature it might. Talk to your company's help desk.

Good luck,

Morris

 

[Zavar]

I've been experimenting with QOS and my work laptop AnyConnect VPN. I definitely get better performance on my work laptop when QOS is disabled. However, other devices on the network can sometimes suffer, so the overall home network is definitely worse when QOS isn't on. The companies IT Admin will not enable split tunneling on the VPN, so no relief that way. I guess I will have to live with slightly degraded VPN performance and keep the rest of the house content!

Thanks everyone for your help!

 

[TonyK132]

I just disabled FreshQoS and re-installed cake. Seems to be working OK, except in the logs I see these lines every 3 minutes. The only thing I see in cru that happens every 3 min is connmon. Is this behavior expected?

Jul 6 07:06:00 RT-AC86U-1BD0 CakeQOS-Merlin: Stopping
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Initialized Runner Unicast Layer
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Initialized Runner Multicast Layer
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: [0;36;44mBroadcom Packet Flow Cache HW acceleration enabled.[0m
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Enabled Runner binding to Flow Cache
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Initialized Runner Protocol Layer (800)
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Broadcom Runner Blog Driver Char Driver v0.1 Registered <3009>
Jul 6 07:06:09 RT-AC86U-1BD0 kernel: [JFFS2 DBG] (7003) jffs2_sum_write_data: Not enough space for summary, padsize = -7
Jul 6 07:07:02 RT-AC86U-1BD0 CakeQOS-Merlin: Starting - ( 100Mbit | 4.5Mbit | besteffort | docsis | docsis ack-filter )
Jul 6 07:07:02 RT-AC86U-1BD0 kernel: [0;33;45mBroadcom Packet Flow Cache HW acceleration disabled.[0m
Jul 6 07:07:02 RT-AC86U-1BD0 kernel: Disabled Runner binding to Flow Cache
Jul 6 07:09:01 RT-AC86U-1BD0 CakeQOS-Merlin: Stopping

 

[dave14305]

I just disabled FreshQoS and re-installed cake. Seems to be working OK, except in the logs I see these lines every 3 minutes. The only thing I see in cru that happens every 3 min is connmon. Is this behavior expected?

Jul 6 07:06:00 RT-AC86U-1BD0 CakeQOS-Merlin: Stopping
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Initialized Runner Unicast Layer
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Initialized Runner Multicast Layer
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: [0;36;44mBroadcom Packet Flow Cache HW acceleration enabled.[0m
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Enabled Runner binding to Flow Cache
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Initialized Runner Protocol Layer (800)
Jul 6 07:06:00 RT-AC86U-1BD0 kernel: Broadcom Runner Blog Driver Char Driver v0.1 Registered <3009>
Jul 6 07:06:09 RT-AC86U-1BD0 kernel: [JFFS2 DBG] (7003) jffs2_sum_write_data: Not enough space for summary, padsize = -7
Jul 6 07:07:02 RT-AC86U-1BD0 CakeQOS-Merlin: Starting - ( 100Mbit | 4.5Mbit | besteffort | docsis | docsis ack-filter )
Jul 6 07:07:02 RT-AC86U-1BD0 kernel: [0;33;45mBroadcom Packet Flow Cache HW acceleration disabled.[0m
Jul 6 07:07:02 RT-AC86U-1BD0 kernel: Disabled Runner binding to Flow Cache
Jul 6 07:09:01 RT-AC86U-1BD0 CakeQOS-Merlin: Stopping

You’re not running version 2.x if you see those Cake…Stopping messages.

 

[Jack Yaz]

You’re not running version 2.x if you see those Cake…Stopping messages.

check the f/w of that user, last i checked they refused to update to 386

 

[dave14305]

check the f/w of that user, last i checked they refused to update to 386

Mystery solved. Old firmware, old script.

 

[TonyK132]

Cake=1.0.7. So, this is a problem that got solved in 386 and cake 2.0?

 

[dave14305]

Cake=1.0.7. So, this is a problem that got solved in 386 and cake 2.0?

It’s not really a problem, but different behavior when Cake wasn’t part of the firmware. When the old script would stop Cake, it would enable all the HW acceleration again. And disable it when Cake was started.

 

[TonyK132]

OK, but does it make sense for the connmon script to cause cake to stop only for it to start 1 min later?

 

[dave14305]

OK, but does it make sense for the connmon script to cause cake to stop only for it to start 1 min later?

That’s really a question for the connmon thread since it is the script triggering the stoppage. But it is optional, AFAIK with the “Exclude ping tests from QoS” option.

 

[TonyK132]

Thanks. I'll post this issue over there. BTW, I did have it set to exclude ping tests from QoS. I just reset it, so after a while, I'll see what it does.

 

[Jack Yaz]

Thanks. I'll post this issue over there. BTW, I did have it set to exclude ping tests from QoS. I just reset it, so after a while, I'll see what it does.

while specific to connmon/spdmerlin, this info may be useful to other cake users. you'll need to disable exclusion from qos. excluding from cake is only possible by turning cake off, unlike adaptive qos where firewall rules can be used to exclude it. therefore, by enabling the option to exclude from qos, you'll see cake turned off and on whenever a test runs

 

[dave14305]

while specific to connmon/spdmerlin, this info may be useful to other cake users. you'll need to disable exclusion from qos. excluding from cake is only possible by turning cake off, unlike adaptive qos where firewall rules can be used to exclude it. therefore, by enabling the option to exclude from qos, you'll see cake turned off and on whenever a test runs

Just for conversation, was there a problem with pings when Cake or other QoS qdiscs were active? I can see doing it for the purposes of speedtests, but pings are supposed to be the litmus test of effective QoS/AQM.

 

[Jack Yaz]

Just for conversation, was there a problem with pings when Cake or other QoS qdiscs were active? I can see doing it for the purposes of speedtests, but pings are supposed to be the litmus test of effective QoS/AQM.

Not that I know of, but i think someone requested it as an option and I couldn't think of a reason not to. I've possibly got the default the "wrong" way round!

 

[TonyK132]

Since saying No to Exclude ping tests from QoS, cake has not been disabled every 3 min like before, so thanks to Jack and Dave for their help in understanding this issue.

Jack - I would vote for making the default "No". The connmon stats do not appear to be impacted in any way by changing that option to No. In fact, JMTC, I would vote for eliminating the option, unless 386 and cake 2.0 does not have the issue.

 

[Zavar]

@dave14305 Haven't seen much regarding Beta 2.1, can you give us an update on progress and if you have a potential ETA for when it will be released to the stable branch?

 

[dave14305]

@dave14305 Haven't seen much regarding Beta 2.1, can you give us an update on progress and if you have a potential ETA for when it will be released to the stable branch?

I've not been doing any programming lately, so there's nothing happening at all to my knowledge. I'm just a contributor to the project. ttgapers is the owner.

Is the beta version working for you?

 

[Zavar]

I haven't been using it, was waiting for it to be a final release.

 

[ttgapers]

I've not been doing any programming lately, so there's nothing happening at all to my knowledge. I'm just a contributor to the project. ttgapers is the owner.

Is the beta version working for you?

Hi all,

This is a question being asked prior to promotion. I have been using it exclusively for quite some time now and through firmware upgrades. If there are others currently running beta 2.1 without issues I/we would be more than happy to promote it to PROD.

Awaiting feedback -- but I would say 2.1 is stable as-is.

 

[pdc]

I've been running it for so long I forgot it was beta

No problems at all, I'd say ship it!