Asuswrt-Merlin 378.55 Beta is now available

[RMerlin]

There's a coding flaw in AdaptiveQos_Bandwidth_Monitor.asp that causes it to error out. Sometimes it'll work just fine (such as right after you reset your router), other times, it'll fail out. I've noticed it most often after I've gone through and named all the devices on the network. Unfortunately, I'm not familiar enough with the code yet to explain why that has anything to do it with it.

I wrote up a summary on it earlier in the week, but didn't post it here (emailed since it included images and such).

Asus has done a number of fixes to that page in upstream code, I'm waiting for a new GPL to merge their changes back. That's why I didn't spend any time tracking down the issue on the current version.

 

[RMerlin]

I see in your sig you are using one of the Seagate 5TB externals. Some people have been reporting slow downs and connection drops on those drives when doing extended writes. If you do a search on 'Seagate dropped connections' you'll find some of the posts.

I've never had anything but problems with the various Seagate devices I've sold/setup for my customers, both internal and external. Internal ones tend to die quite early, regardless of the model (so no, it's not just limited to the infamous 7200.11 model). As for external, various customers have been tons of reliability/stability issues with various external Seagate hard drives models. Even when plugged directly to a computer these would frequently disconnect and reconnect while under use. The USB enclosure they use is simply garbage, they love to use proprietary USB cables, and if the disk does not come with an AC adapter you can be sure that this disk will never work reliably regardless on where you plug it. On at least one case I ended up scrapping the Seagate enclosure, and moving the disk to a Vantec enclosure, which has been working flawlessly storing server backups since then.

My advice is to avoid any Seagate hard disk like the plague if you value stability. Get an OEM Western Digital or Hitachi disk, and put it inside a Vantec enclosure. External WD devices might also be good, I just don't have much first-hand experience with these so I can't comment.

There's a reason why those external Seagates are regularly discounted at Best Buy/FutureShop/NCIX.

 

[RMerlin]

Traffic Analyzer - Statistic Inaccurate

Closed source, out of my control.

 

[RMerlin]

I tested the two topic above:
a) The new login has a major disadvantage from my point of view
- The new login page does not support browser autocomplete functionalty for the login credentials (at least in Firefox v39).

Asus has deliberately disabled that on the new login page (they actually put code in place to prevent autofilling from working). Since this page is very new, and I already see a lot of changes to it in newer beta firmware, I'm waiting for it to stabilize before evaluating what needs to be changed to that page. I need to be careful with what I do with that page. As a login page, the primary concern for that page is security.

b) I do not see any change in the memory behavior
- Based on my "special memory configuration for Transmission" (see details here), I experiance the same nice "bumpy road" in memory allocation and automatic relase.
- Is this due to my configuration or in general the memory is still release periodically on other heavy used installations?

You are already actively doing disk I/O with your torrent client, so most likely your cache gets repopulated before you even notice it got flushed. So for that specific scenario, the memory fix has no effect. Torrenting from a router with 256 MB of RAM and low I/O performance storage is simply a huge strain on resources.

 

[RMerlin]

The new wireless time scheduler does not work for me. If I select enable and then select the allowed times the WiFi just switches off regardless after a minute or so. The system time is correct and I have tried applying disable in the scheduler too (in case they had been inadvertantly swapped in the code) but WiFi still switches off.

The time scheduler is nothing new, it's been part of Asuswrt for years. I haven't seen any change to that code from Asus in recent months.

 

[mromero]

The exception I think is the Seagate Backup Plus Slim that actually has a great Saumsung drive inside it.

I've never had anything but problems with the various Seagate devices I've sold/setup for my customers, both internal and external.

 

[RMerlin]

The exception I think is the Seagate Backup Plus S-l-i-m that actually has a great Saumsung drive inside it.

It's a portable enclosure, so I bet it doesn't provide enough power for stable operations.

(edit: removed keyword that's probably triggering spam filter)

 

[win465]

Thank you Merlin, this is a much better beta than the latest Asus 87u one. I have one question to ask. Why when I install a new firmware my router seems super fast, but then after a day it fells sluggish?

 

[GoNz0]

any ac-68u users running dnscrypt on this build? mine no longer works.

 

[HeadBanger]

The time scheduler is nothing new, it's been part of Asuswrt for years. I haven't seen any change to that code from Asus in recent months.

Merlin you clearly don't use it as it is completely different for this release!!

 

[RussellInCincinnati]

The USB enclosure they use is simply garbage, they love to use proprietary USB cables, and if the disk does not come with an AC adapter you can be sure that this disk will never work reliably regardless on where you plug it. On at least one case I ended up scrapping the Seagate enclosure, and moving the disk to a Vantec enclosure...

The Inateck enclosures are also nice, with use of way above average chipsets and the like. Inateck unfortunately does not make a 3.5 inch hard drive enclosure per se, but they make plenty of different 3.5 inch-compatible hard drive docking stations. Of course they make lots of 2.5 enclosures but that probably is irrelevant to the original poster's 5 terabyte drive.

 

[mromero]

It's a portable enclosure, so I bet it doesn't provide enough power for stable operations.

(edit: removed keyword that's probably triggering spam filter)

This is a USB 3.0 drive and plugged into the AC RT87U works like a champ. Even on an underpowered Raspberry Pi narry a hiccup. The cable has no additional connector to get extra power such as you see on other external HDs and does not need it. Many gamers get this drive, open it, remove the HD, and use it on the PS4s.

 

[owine]

Upgraded to b1 and my dnsmasq.conf.add file caused issues during boot multiple times. By renaming the file I was able to boot successfully and had no issues naming the file back and restarting dnsmasq. Will keep an eye on it in future reboots.

No issues for me saving or filling the login form with 1Password on my Mac.

 

[Gitsum]

Noticed 100% CPU usage on one core of my AC87U when downloading a torrent at my connections full speed of 65mbps. CTF is on. QOS is on. App analysis is on.

 

[CivHunter]

I have installed 378.55_beta1 over 378.54_1 on my RT-AC87R, no default resets done, and so far I notice two things. Firstly my Galaxy Note 4 is not disconnecting and reconnecting on the 5ghz channel anymore (although it worked ok with the last firmware but was always reconnecting.) Secondly when I log into the router remotely and open the AdaptiveQoS_Bandwidth_Monitor.asp page I am seeing no indication of data usage and I can not expand the view on any of my devices to see what apps are using data. I also can not reorder devices or drag and drop the priority tags to reassign them priorities. Its like the table is static only now.
I am using chrome for my browser and I can click on the priorities and filter my list of devices but that's about it.

New discovery on the adaptive QOS page. Using the browser on my Galaxy Note 4 the pages work correctly IF I login to the Web ui while connected to WiFi but is still non-functional if I am logging in remotely from the WAN. (No reboots or resets where done to the router besides the original flash to this new beta.)

 

[RMerlin]

Merlin you clearly don't use it as it is completely different for this release!!

I see they switched to the same webui layout they were using for Parental Control now. Not sure when the change actually occurred, the old UI was hidden by default, and I never saw the change mentioned in their changelog.

I just configured a schedule, which correctly disabled wifi. Will know at the next hour if wifi gets properly turned on as scheduled.

 

[RMerlin]

Asus rewrote the time scheduling code, and I'm getting lost in the multiple arrays they are simultaneously juggling with there (with no comments as to how the code works):

https://github.com/RMerl/asuswrt-merlin/blob/master/release/src/router/rc/watchdog.c#L1007

The start time properly gets handled, but it fails to properly set the end time of a given time period. So once a period is set to "enable", the entire rest of the week is set to enable as well.

I give up on figuring out this code for now, and will wait their next GPL release in hopes it contains a fix.

 

[HeadBanger]

OK thanks for the confirmation. Is the old working time scheduler completely gone or just hidden? Is there a way to get to it if so?

 

[Martineau]

Asus has deliberately disabled that on the new login page (they actually put code in place to prevent autofilling from working). Since this page is very new, and I already see a lot of changes to it in newer beta firmware, I'm waiting for it to stabilize before evaluating what needs to be changed to that page. I need to be careful with what I do with that page. As a login page, the primary concern for that page is security.

Not quite sure what you would change to the new logon process? - apart from already fixing the scaling issue to make the Logon fields visible/accessible on your screen!
As a security measure many logon forms (Citrix in particular) explicitly prevent the use of cut'n'paste and this can be a real pain when the helpdesk ping you with a temporary reset password that is very strong such as $SwG60OI1!. (includes the period) that you must carefully type blind into the password field

I prefer to use scripting and I haven't had to manually enter my credentials to access an ASUS router for years!
However, I acknowledge that the 'art' of scripting logons is potentially vulnerable, as passing the router credentials as a command line argument (or even extracting them from an obfuscated Registry key) to the script at some point means the password may be disclosed in plain text - unlike commercial secure password vaulting such as LastPass etc.?

Furthermore, scripting allows more than a simple logon..
My example physically logs onto the router and then auto-navigates to display the Wireless log.

I'm sure as a developer you probably have a set of similar scripts to eliminate the tedious manual keystrokes needed to test new GUI features etc.?

http://tinypic.com/r/24eb0vn/8

 

[GoNz0]

autofill works fine using lastpass