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

[john9527]

That led me to this which exactly describes my problem. So I'm not sure there's a satisfactory solution at the moment.

Nice find. Can you try 2s and 4s as well?

EDIT: There must be a server or routing contributor here as well. With the default 10s idle-timeout, I'd see an occasional spurious switch, but nothing like what you were seeing.

 

[ColinTaylor]

Nice find. Can you try 2s and 4s as well?

OK. I'll report back.

 

[john9527]

@ColinTaylor

Let's start with you trying 5s first. I'm thinking about changing the default to 5s, and then add an nvram var to allow the user to change it pending the update they mentioned in the issue thread.

 

[jrmwvu04]

Given the pattern of activity (even though it's rather vague) my gut feeling is that it's an issue with the upstream servers not responding quickly enough (possibly due to load). I suspect 9.9.9.9 is much worse than 1.1.1.1 in this respect.

I’m finding the same to be true. Not as much scientific data to back it up as you, but the user experience is just generally smoother.

 

[ColinTaylor]

Let's start with you trying 5s first. I'm thinking about changing the default to 5s, and then add an nvram var to allow the user to change it pending the update they mentioned in the issue thread.

OK I've only been testing for a few minutes but one thing is immediately apparent. With 9.9.9.9 as the primary and a 4 second timeout, when stubby starts it connects once to 9.9.9.9 after which it switches to 1.1.1.1. It stayed on 1.1.1.1 for a few minutes until I terminated the test. Reversing the order of servers and restarting stubby had the connection not switch from 1.1.1.1 at all (for the few minutes of the test).

Changing the timeout to 2 seconds with 9.9.9.9 as the primary saw it stick with 9.9.9.9 for the first 4 connections before switching to 1.1.1.1 (and it's still there after 7 minutes).

 

[john9527]

OK I've only been testing for a few minutes but one thing is immediately apparent.

What you describe is the behavior I see at 10s.....so it's definitely location dependent.

Can you try this set of changes?

pc_insert "tls_query_padding_blocksize" "tls_connection_retries: 5" $CONFIG
pc_insert "idle_timeout" "timeout: 2000" $CONFIG
pc_replace "idle_timeout: 10000" "idle_timeout: 2000" $CONFIG

 

[ColinTaylor]

What you describe is the behavior I see at 10s.....so it's definitely location dependent.

Can you try this set of changes?

pc_insert "tls_query_padding_blocksize" "tls_connection_retries: 5" $CONFIG
pc_insert "idle_timeout" "timeout: 2000" $CONFIG
pc_replace "idle_timeout: 10000" "idle_timeout: 2000" $CONFIG

OK Trying it now. BTW What's the name/directory for the postconf script?

 

[john9527]

BTW What's the name/directory for the postconf script?

/jffs/scripts/stubby.postconf (same as the other postconf scripts)

it modifies /etc/stubby.yml

 

[ColinTaylor]

@john9527 OK this is looking promising. Been running for 20 minutes now. The time before the first Conn_shuts is about the same as it was previously when set to 2 seconds, but this time it's not switching servers it's just reconnecting to 9.9.9.9.

[00:35:21.616360] STUBBY: 9.9.9.9                                  : Conn closed: TLS - Resps=     1, Timeouts  =     0, Curr_auth =Success, Keepalive(ms)=  2000
[00:35:21.616547] STUBBY: 9.9.9.9                                  : Upstream   : TLS - Resps=   166, Timeouts  =     0, Best_auth =Success
[00:35:21.616669] STUBBY: 9.9.9.9                                  : Upstream   : TLS - Conns=    60, Conn_fails=     0, Conn_shuts=      5, Backoffs     =     0

 

[Dikij]

Hi Everybody, maybe you Gays add function AP Repeater 802.1x with EAP-PEAP-mschapv2 authorization.

In Europe big ISP (“UPC” his name) have many free access point in Europe. But ASUS firmware not support EAP AP. Maybe in this fork you have attach EAP for access. Please help!

 

[loveleeyoungae]

I have a N66U running Merlin 374.41. Can I use the WebGUI? (As the UpgradeMatrix table suggests to use the Restoration Tool).

EDIT: So, see the UpgradeMatrix mentioned about V28. So browsed in V27 and earlier versions, I didn't see that UpgradeMatrix file. Does it mean I can upgrade to V27 via WebGUI?
(Yeah, I can take time to upgrade to V27, then continue to V34. Don't like the ASUS Restoration Tool much).

 

[ColinTaylor]

@john9527 Uptime of 11 and a half hours. No switches have occurred despite having experienced 24 timeouts.

[11:32:58.094953] STUBBY: 9.9.9.9                                  : Conn closed: TLS - Resps=     1, Timeouts  =     0, Curr_auth =Success, Keepalive(ms)=  2000
[11:32:58.095143] STUBBY: 9.9.9.9                                  : Upstream   : TLS - Resps=  3242, Timeouts  =    24, Best_auth =Success
[11:32:58.095264] STUBBY: 9.9.9.9                                  : Upstream   : TLS - Conns=  1020, Conn_fails=     0, Conn_shuts=    106, Backoffs     =     0

 

[john9527]

I have a N66U running Merlin 374.41. Can I use the WebGUI? (As the UpgradeMatrix table suggests to use the Restoration Tool).

EDIT: So, see the UpgradeMatrix mentioned about V28. So browsed in V27 and earlier versions, I didn't see that UpgradeMatrix file. Does it mean I can upgrade to V27 via WebGUI?
(Yeah, I can take time to upgrade to V27, then continue to V34. Don't like the ASUS Restoration Tool much).

You should be able to go directly to the latest release via the gui, but since it's such a big jump I would make sure to do a factory reset following the upgrade, select to do a 'format jffs at next boot', then reboot twice.
Next step then is to reconfigure with your desired settings (don't use a saved configuration).

 

[john9527]

That led me to this which exactly describes my problem. So I'm not sure there's a satisfactory solution at the moment.

FYI....I added a note to the issue saying it was found on my fork and offering to help test any new releases.

Uptime of 11 and a half hours. No switches have occurred despite having experienced 24 timeouts.

Good feedback! The 35T1 test release includes these latest set of changes.

 

[loveleeyoungae]

You should be able to go directly to the latest release via the gui, but since it's such a big jump I would make sure to do a factory reset following the upgrade, select to do a 'format jffs at next boot', then reboot twice.
Next step then is to reconfigure with your desired settings (don't use a saved configuration).

Got some issues and took some time to resolve them. I'll list them here, maybe it's good to add to installation notes for other people to know.
- The direct upgrade via WebGUI from Merlin 374.41 to v34E3 was successful. Things seemed good: Internet was available, WebGUI access works normally, old settings were retained. However, I did not test more as I proceeded to perform the Factory Defaults right away. That's when issues arose.
- Right when the router came back, it welcomed me with a quite slow QIS WebGUI. Did not know how I luckily clicked past those QIS setup pages, usually had to click 2-3 times, when a page loaded, it usually loaded without images just like when you browsed a slow page and their images were timed out).
- After the QIS Welcome, the WebGUI pages showed all web elements (images, buttons) normally, but were still slow. Clicking every function needed around 15-30s for a page to load.
- Alright, maybe because there was still one thing we hadn't done yet: format JFFS. So I applied the "Format JFFS at next boot", and clicked Reboot.
- The WebGUI was then still slow. Tried some more Factory Defaults, Format JFFS and Reboot. WebGUI was still slow.
- Thought maybe something with the nvram (?!). So I ssh'ed and performed the "mtd-erase -d nvram".
- Alright, the QIS Welcome showed up fast, webGUI worked normally. Thought things were resolved. But no, Internet was not available (Actually, it was the laptop wired to N66U did not have Internet connection, I did not check on the router itself which I'll mention in later steps).
ISP Modem ========= WAN[Router]LAN ==============Laptop
- Performed the Factory Defaults, and webGUI became slow again, and this time it was worse as the laptop still did not have Internet connection.
- Alright, maybe it was time to perform an upgrade using Restoration Tool or CFE miniWeb. I tried the CFE miniWeb method.
- I *skipped* the QIS Welcome and browsed 192.168.1.1/index.asp directly as I was so fed up with that step in previous tries. Terrible mistake. I saw that the router had Internet connection, on Network tools I could ping websites, but the laptop still did not have. Did not know the reason, I tried another chaos of Factory Defaults, clear nvram using WPS button... and always *skipped* QIS Welcome, so the issue was always there.
- Ok, held my breath, tried everything from start again: CFE miniWeb, QIS Welcome page showed up fast for me to perform inital setup, Factory Defaults, went through QIS Welcome page once again, applied "Format JFFS" right away, reboot, reboot again. Now everything is fine.

In summary, here are my thoughts on the order of the steps to do for a successful upgrade from Merlin 374.41 to 34E3:
- Direct upgrading via webGUI works, but obviously normal Factory Defaults cause issues for the webGUI.
- So, instead of perfoming Factory Defaults after upgrading, you would want to clear nvram via SSH first (maybe better than WPS method?)
- Only after that, QIS Welcome page would show up fast for you to perform initial setup. And do not skip that step.
- Now you can proceed to apply Format JFFS. Click Reboot.
- Finally, perform Factory Defaults. Wait for the router to reboot.
- Go through QIS Welcome again. You're done.


Still curious why QIS Welcome is a must, I have always thought it is just a wizard so we can skip and set things up manually...?

 

[JWoo]

How can we get the 35T1 test release that includes the latest set of changes for DNS over TLS. Thank you!

 

[RMerlin]

How can we get the 35T1 test release that includes the latest set of changes for DNS over TLS. Thank you!

https://www.snbforums.com/threads/test-asuswrt-merlin-lts-fork-multiple-items.48642

 

[john9527]

- So, instead of perfoming Factory Defaults after upgrading, you would want to clear nvram via SSH first (maybe better than WPS method?)

I must admin, I'm baffled....as doing an nvram erase followed by a reboot is exactly the same as a factory default reset.

The web gui loading without images is usually a symptom of a mismatch in the browser cache or with the HTTPS cert (many browsers will change HTTP access to HTTPS automatically if you used HTTPS in the past....drives me crazy!). So I think the missing step would be to clear your browser cache/cookies.

 

[Techneaux]

I'm having an issue with a wifi delay after two android (Google Pixel) phones go to wifi powersave mode (I think, not sure, see below).

Upgraded N66U from an old merlin build to latest 34E3 using restore mode. Factory reset afterwards.
(These two android phones are the only mobile devices I have to test.)

Only sometimes after either phone has not been used for >10 minutes and turned back on,

• wifi is listed as "connected" on the android phone itself (icon on top is still connected icon - not sure how much that means)
• trying to browse in chrome app to a website: website doesnt load, eventually get a "err name not resolved"
• after 10-20 second delay, works again and website loads

No problems with wifi dropouts on any other devices (e.g. desktop PCs). Not sure if root cause is phone-specific or from new 34E3 build.
I tried switching between DNS servers too (ISP and cloudflare).
Difficult to test because only sometimes this problem occurs... perhaps wifi powersave is not active the other times.

Any suggestions? Is there a specific logging level I should turn on to gain more insight?
Thank you!

 

[jrmwvu04]

The web gui loading without images is usually a symptom of a mismatch in the browser cache or with the HTTPS cert (many browsers will change HTTP access to HTTPS automatically if you used HTTPS in the past....drives me crazy!). So I think the missing step would be to clear your browser cache/cookies.

Definitely. I went through very similar setup issues many, many moons ago and I learned two things about doing the setup.

One, clearing out browser cache frequently is a very helpful step when you’re changing settings, let alone doing resets. The usefulness probably depends on the browser but safari for instance loves to cache everything and it can cause problems.

Two, and I’m much less clear on this, so I’ll use all my personal anecdotes - doing router factory resets seems to confuse the Comcast connection. Power cycling the modem in tandem seemed to help, but one thing I noticed is that the Routers mac address setting in wan settings (the one with the button to clone the address) was sometimes populated. Which caused problems for whatever reason. Clearing it seemed to be required for things to work right. And doing the QIS seemed to mess with that, setting it and causing problems.