[Fork] Asuswrt-Merlin 374.43 LTS releases (Archive)

[TekWarren]

@john9527 I tagged you in another thread about an unrelated issue but have a question: Is there supposed to be an option to allow enabling/disabling of "Traffic Analyzer"? This is being suggested to me as a way to improve VPN throughput on my AC66U.

Thanks

 

[john9527]

@john9527 I tagged you in another thread about an unrelated issue but have a question: Is there supposed to be an option to allow enabling/disabling of "Traffic Analyzer"? This is being suggested to me as a way to improve VPN throughput on my AC66U.

Thanks

Traffic Analyzer is part of the Trend Micro DPI engine.....two things....
- The Trend Micro engine isn't part of the fork (came later)
- Even if it was, it's not available for the AC66U (only supported on ARM routers in Merlin/Asus fw)

Sorry....

 

[Aayman F]

Hi RTMerlin, I just signed up to ask you a question.

My old RT-N16 with Tomato died recently and I upgraded to the RTN66U.

I flashed it with your FW (amazing work btw) since I wanted something simple and not as advanced as Tomato.

However I did notice that your QoS functions differently from Tomato.

My setup:
Connection: Comacast 30Mb|6Mb

Modem - cable to desktop gaming PC AND cable to RT N66U (for everything else)

QoS Setting - 25Mb|4.5Mb

On my old router, the setup functioned as if my connection was 25Mb|4.5Mb and I never experienced ping spikes or lag on my desktop with multiple devices connected to Wifi

I tried the same configuration with your FW and ran a test from speedtest.net while pinging google from my desktop. While the download is tested, the router lets it go upto 30Mb (ignoring 25Mb) and I get high ping on the desktop, but the router obeys the 4.5Mb setting and I don't get high ping on the desktop.

Essentially, 30Mb|4.5Mb with one wifi device connected.

I think I read somewhere that when only one device is connected, it should be able to use all the bandwidth, but QoS kicks in when multiple devices connect. But the upload bandwidth is restricted even when only one device is connected. Is there anyway that I can force it to obey the download limit as well?

Sorry for the long post, I just want you to have a complete idea of what's going on.

 

[john9527]

Hi RTMerlin, I just signed up to ask you a question.

My old RT-N16 with Tomato died recently and I upgraded to the RTN66U.

I flashed it with your FW (amazing work btw) since I wanted something simple and not as advanced as Tomato.

However I did notice that your QoS functions differently from Tomato.

My setup:
Connection: Comacast 30Mb|6Mb

Modem - cable to desktop gaming PC AND cable to RT N66U (for everything else)

QoS Setting - 25Mb|4.5Mb

On my old router, the setup functioned as if my connection was 25Mb|4.5Mb and I never experienced ping spikes or lag on my desktop with multiple devices connected to Wifi

I tried the same configuration with your FW and ran a test from speedtest.net while pinging google from my desktop. While the download is tested, the router lets it go upto 30Mb (ignoring 25Mb) and I get high ping on the desktop, but the router obeys the 4.5Mb setting and I don't get high ping on the desktop.

Essentially, 30Mb|4.5Mb with one wifi device connected.

I think I read somewhere that when only one device is connected, it should be able to use all the bandwidth, but QoS kicks in when multiple devices connect. But the upload bandwidth is restricted even when only one device is connected. Is there anyway that I can force it to obey the download limit as well?

Sorry for the long post, I just want you to have a complete idea of what's going on.

First, welcome to snbforums....Now to your questions....

Just for clarification, this firmware is a fork based on Merlin's release from about a year ago that I am now maintaining, not Merlin. There are some reasons why some people prefer this level, as detailed in the first post in this thread. If you were looking for the latest Merlin, the link can be found in another 'sticky' near the top of this forum.

Now to QoS....what you describe is the way the current 'traditional QoS' implementation works in all the Asus/Merlin levels. There is not a way to do bandwidth limiting either globally or by client. Asus is currently rolling out some beta levels with a bandwidth limiter function, but I'm not sure if it's going to be available on all models or just the ARM based routers (AC56 and above), so you should watch the forums.

I've also been working on a global bandwidth limiter, but have been waiting to see what Asus is going to do. I might release that as a beta level in the near future.

There's a Tomato for the N66, so you may need to go back to that (at least for a while) for QoS to work as before.

 

[Just Checking]

@Just Checking - One of the things that struck me is you mentioned a 10dBm change....I think that's over an 8x reduction in power...very significant. Just thinking out loud.....could your Elitebook been on battery vs plugged in when comparing measurements?

EDIT: Another thought just to verify....you are using fixed channels for the measurements.

Yes, a 10dBi change is very significant. That is why I was very concerned.

I try to do all of my measurements as close to the same as possible and I know that if I do not have a good battery, the readings will be different. The Elitebook I use for testing has a good, extended life battery which is as close to fully charged as possible when I do my testing and firmware updates (I get paranoid about losing connections or power during a FW flash and bricking a router). In this case I was plugged in and using the charger power.

As an update, I downloaded the ver 14 FW again and erased the old downloads. I checked that the file size was the same and they they were. I then set up and tested the signal strength on both 2.4GHz and 5GHz bands for an RT-AC66 router that I had previously had the signal strength loss when going from ver 13E1 to ver 14. I flashed the ver 14 FW and rebooted the router. This time the signal strength was the same with the ver 14 FW vs. the ver 13E1 FW. I tried another RT-AC66 and that one was successful also. I then tried a RT-AC68 router and that one was good too. Finally, I tried to flash the ver 14 fork FW on to a RT-AC68 that I had previously installed the HGGomes Fork FW. That one would not take the flash. I then did a hard reboot to clear the nvram cache and installed RMerlins 378.55_0 FW then did another hard reboot with nvram clear and it would finally take the 374.43_2-14j9527 FW flash. The signal strength did change between the FW types but, since the power settings are so different between the two FW forks, I cannot say if it made any difference at all.

I know that my measurements of signal strength are as precise and reproduced as close as possible between the testing events. I really don't know why the FW flash to ver 14 gave "good" results this time and not the last time. Perhaps I did have a corrupted download the first time. If so, it wasn't an obvious difference. I am also not sure that the ver 14 FW actually took and is running instead of the ver 13E1 that was previously on most of the routers I tested. Yes, I am sure on the one that was running HGGomes FW since it is so different but I can't tell anything about signal strength vs. power setting for that one. It may remain a mystery. At least my problem is solved successfully.

BTW, I switched back from the HGGomes FW on the RT-AC68 because the CPU temps are higher with HGGomes FW than with John's fork FW even though CPU utilization is not noticeably different with either one. Maybe the algorithm that HGGomes uses to calculate the CPU temp is different than with other FW. I noticed that the HGGomes FW indicated higher CPU temps on an RT-AC87 router than RMerlins 378.55_0 FW too so it isn't a router specific difference.

 

[Just Checking]

@Just Checking - One of the things that struck me is you mentioned a 10dBm change....I think that's over an 8x reduction in power...very significant. Just thinking out loud.....could your Elitebook been on battery vs plugged in when comparing measurements?

EDIT: Another thought just to verify....you are using fixed channels for the measurements.

All of the routers I maintain are set on fixed channels to minimize interference with all the other wireless signals in the surrounding environment. I also set the channel width to 20 for the 2.4GHz signal and to 80 for the 5GHz signal.

 

[john9527]

A quick update to address some issues...Thanks to everyone for providing feedback.

LATEST RELEASE: Update-14E1
17-September-2015
Merlin fork 374.43_2-14E1j9527
Download http://1drv.ms/1uChm3J
============================

Update-14E1 HighLights:

• Fix for not being able to access the gui after a factory reset - reported by @endlesszeal
• Fix for setting the correct path when executing remote ssh/plink commands - reported by @000111, @Ren34
• Fixes for improved handling of http port change
• Fix for networkmap intermittently not showing correct client type

A factory reset is not required coming from any previous fork level.

And, for those who like to try new things....
There is also an experimental option to improve QOS (bufferbloat), based on input from @Chrysalis
This option sets SFQ packet limits through an algorithm based on your upload bandwidth. The option is disabled by default (same as previous releases) and is enabled via telnet or SSH logon to the router.

nvram set qos_sfql=[0=default, 1=auto, 2-127 set sfq packet limit] - that's a small 'L' in qos_sfql
nvram commit

followed by a reboot or a restart of QOS. The algorithm was developed on results from my connection only, so I'm interested in how it may perform for others. Feedback is welcome.

Enjoy!

 

[john9527]

This time the signal strength was the same with the ver 14 FW vs. the ver 13E1 FW.

Glad to hear this went into the category of 'Unsolved Mysteries'. It seems like you have a very methodical test method which is excellent. That's why I couldn't understand if it was generally true, why there wasn't a flood of 'What happened?" messages.

 

[Just Checking]

Glad to hear this went into the category of 'Unsolved Mysteries'. It seems like you have a very methodical test method which is excellent. That's why I couldn't understand if it was generally true, why there wasn't a flood of 'What happened?" messages.

Thanks for the FW John! Having FW that makes these routers reliable and fully featured makes my life much easier, and my clients happy!

 

[SuperFly74]

Cant crack the TX op to more then 100mW ? Why


Sent from my iPhone using Tapatalk

 

[SuperFly74]

Can u hide your ip VPN with fork


Sent from my iPhone using Tapatalk

 

[telUK]

hi

most likely being a plonker here, just upgraded from v13 to v14e1, noticed time zone was set to paris after a nvram reset,

Looking under timezones I have greenwich mean time and greenwich mean time Edinburgh, Lisbon , London.

Being in London what should I select?

thanks

 

[sbailey4]

hi

most likely being a plonker here, just upgraded from v13 to v14e1, noticed time zone was set to paris after a nvram reset,

Looking under timezones I have greenwich mean time and greenwich mean time Edinburgh, Lisbon , London.

Being in London what should I select?

thanks

Well if it were me I would select London and see what time appears. If its incorrect select the other one

 

[john9527]

hi

most likely being a plonker here, just upgraded from v13 to v14e1, noticed time zone was set to paris after a nvram reset,

Looking under timezones I have greenwich mean time and greenwich mean time Edinburgh, Lisbon , London.

Being in London what should I select?

thanks

They are both GMT....the one with the London label also includes a DST setting (which is probably wrong, since DST has changed so much over the last year). If you choose the DST enabled selection, you'll probably need to check/adjust the DST start/end dates.

 

[ColinTaylor]

Looking under timezones I have greenwich mean time and greenwich mean time Edinburgh, Lisbon , London.

I don't know for sure, but I've always assumed that GMT meant GMT, whereas GMT Edinburgh, Lisbon , London would be GMT in winter and BST in summer.

Doh! John beat me to it.

 

[telUK]

They are both GMT....the one with the London label also includes a DST setting (which is probably wrong, since DST has changed so much over the last year). If you choose the DST enabled selection, you'll probably need to check/adjust the DST start/end dates.

Hi

I changed it to the GMT selection and my time was an hour behind so I changed it back to GMT london,lisbon......

Is It important to have the correct time anyway? I just use my router as just that, I don't use any of the other features on it.

thanks

/edit - 14E1 working well, thanks very much!!

 

[Fuzzy_Dunlop]

This is on an RT-AC68R. When I connect via SSH with a forwarded port for a SOCKS proxy, the SSH connection immediately drops when I attempt to browse to any site on Chrome via the SOCKS proxy. Connecting to my desktop via VNC over an SSH tunnel seems to work fine and doesn't disconnect it.

I have previously done this on all previous Merlin releases as well as with Tomato on my old RT-N16 without an issue.

I originally had this problem with v14 (first release I have installed), but thought it may be fixed with 14E1 so I waited to report it.

Edit: When I switched from the newer Merlin branch I reset all defaults and restored all settings via the Merlin settings restore script.

 

[kolodzij]

A quick update to address some issues...Thanks to everyone for providing feedback.

LATEST RELEASE: Update-14E1
17-September-2015
Merlin fork 374.43_2-14E1j9527
Download http://1drv.ms/1uChm3J
============================

Update-14E1 HighLights:

• Fix for not being able to access the gui after a factory reset - reported by @endlesszeal
• Fix for setting the correct path when executing remote ssh/plink commands - reported by @000111, @Ren34
• Fixes for improved handling of http port change
• Fix for networkmap intermittently not showing correct client type

A factory reset is not required coming from any previous fork level.

And, for those who like to try new things....
There is also an experimental option to improve QOS (bufferbloat), based on input from @Chrysalis
This option sets SFQ packet limits through an algorithm based on your upload bandwidth. The option is disabled by default (same as previous releases) and is enabled via telnet or SSH logon to the router.

nvram set qos_sfql=[0=default, 1=auto, 2-127 set sfq packet limit] - that's a small 'L' in qos_sfql
nvram commit

followed by a reboot or a restart of QOS. The algorithm was developed on results from my connection only, so I'm interested in how it may perform for others. Feedback is welcome.

Enjoy!

Just loaded V14E1 and everything working great as always.

Some bufferbloat feedback. I have in the past tried to improve scores without much success. I suspect windows device drivers are part of the problem.

with qos_sfql=1 auto

Windows 8.1 desktop with Realtek gigabit wired adapter Bufferbloat grade goes from F to C (nice improvement)

Windows 10 laptop with Intel ac wireless adapter grade goes from B to B

ISP is Cox, Speed tier is 50/5, QOS set to 50/5 and 45/4.5 with same result

Router is RT-AC68P

 

[sbailey4]

BTW John, thanks for keeping this fork going. So far working great. I do simply use the router in basic config as a wireless router. Nothing fancy setup. Signal great, speed great no issues for me so far. Seem as stable as previous versions.

 

[john9527]

This is on an RT-AC68R. When I connect via SSH with a forwarded port for a SOCKS proxy, the SSH connection immediately drops when I attempt to browse to any site on Chrome via the SOCKS proxy. Connecting to my desktop via VNC over an SSH tunnel seems to work fine and doesn't disconnect it.

I have previously done this on all previous Merlin releases as well as with Tomato on my old RT-N16 without an issue.

I originally had this problem with v14 (first release I have installed), but thought it may be fixed with 14E1 so I waited to report it.

Edit: When I switched from the newer Merlin branch I reset all defaults and restored all settings via the Merlin settings restore script.

Thanks for the report. I'm actually running a bit ahead on a couple of components vs Merlin's latest just based on the release cycle.

I've never set up what you describe. Can you provide a bit of detail on 'how-to' so I can try and recreate?