[Release 384/NG] Asuswrt-Merlin 384.5 is now available

[cmkelley]

Off course it's mounted on /dev/sda1 but the AB-SOLUTION folders on it is questionable.

Remove the drive, reboot the router and run fdisk -l again to check if you still see that path.

Also check the services-start and services-stop scripts in /jffs/scripts folder for any rouge entries.

Also post-mount script.

 

[Asad Ali]

Also post-mount script.

Yup forget to add that one. Thanks!!


Sent from my iPhone using Tapatalk

 

[Sebastien Bougie]

Off course it's mounted on /dev/sda1 but the AB-SOLUTION folders on it is questionable.

Remove the drive, reboot the router and run fdisk -l again to check if you still see that path.

Also check the services-start and services-stop scripts in /jffs/scripts folder for any rouge entries.

I think i found where this mounting is from, but i dont know how to fix that. This mapping is with my Debian start. if i unmountit and do a debian restart its bac.. I have checked the S99debian file and the path is :CHROOT_DIR=/mnt/AB-Solution/entware/debian

Is there a way to force this mounting to /dev/sdb5/AB-Solution/entware/debian?

 

[Mark112887]

I'm having an issue that I first started experiencing last night. After approximately 42 days of trouble free up time no reboots no wireless dropouts nothing of any concern, I started noticing a change in my Internet speed. When I did a speed test I noticed that my upload speed was fine exactly what it should be but my download speed was roughly 3.5 megabits per second when regularly it should be between 90 and 100 megabits per second. First thing I did was completely power cycled my modem, the situation did not change. I then called my ISP to check for potential outages, there weren't any. Finally the last thing I did was rebooted the router after which my download speed returned to normal.

I'm running the latest firmware 384.5 on my AC 3200. Has anyone else experienced this? If so do we know what's causing it?

Thanks so much

 

[kfp]

I'm having an issue that I first started experiencing last night. After approximately 42 days of trouble free up time no reboots no wireless dropouts nothing of any concern, I started noticing a change in my Internet speed. When I did a speed test I noticed that my upload speed was fine exactly what it should be but my download speed was roughly 3.5 megabits per second when regularly it should be between 90 and 100 megabits per second. First thing I did was completely power cycled my modem, the situation did not change. I then called my ISP to check for potential outages, there weren't any. Finally the last thing I did was rebooted the router after which my download speed returned to normal.

I'm running the latest firmware 384.5 on my AC 3200. Has anyone else experienced this? If so do we know what's causing it?

Thanks so much

Did you check the CPU/RAM usage before rebooting? Logs? It’s probably really hard to go on with this one off incident without more info.

 

[Mark112887]

Was so tired I totally forgot. If it happens again I'll be sure to save the logs and check RAM/CPU

 

[BlackMage]

when i update to the ng-firmware then HTTPS connections outside stop working(always invalid authentication token)...
on RT-AC68U

 

[kfp]

when i update to the ng-firmware then HTTPS connections outside stop working(always invalid authentication token)...
on RT-AC68U

Assuming you’re coming from a non-ng, have you factory reset and reconfigured?

 

[BlackMage]

Assuming you’re coming from a non-ng, have you factory reset and reconfigured?

Is not there also a possibility to convert the old configuration data?

 

[kfp]

Is not there also a possibility to convert the old configuration data?

Well the fact is you’re having problems now and to be effective in debugging this problem it’s better to start from a known good state.

 

[Asad Ali]

I think i found where this mounting is from, but i dont know how to fix that. This mapping is with my Debian start. if i unmountit and do a debian restart its bac.. I have checked the S99debian file and the path is :CHROOT_DIR=/mnt/AB-Solution/entware/debian

Is there a way to force this mounting to /dev/sdb5/AB-Solution/entware/debian?

I'm not an expert in that case.
Calling @thelonelycoder and other big guys to help.

 

[kfp]

I think i found where this mounting is from, but i dont know how to fix that. This mapping is with my Debian start. if i unmountit and do a debian restart its bac.. I have checked the S99debian file and the path is :CHROOT_DIR=/mnt/AB-Solution/entware/debian

Is there a way to force this mounting to /dev/sdb5/AB-Solution/entware/debian?

So the problem is way earlier than the CHROOT, and that might not be the only place using the wrong path so editing it might not be a good idea. I’m guessing this is what happened:

Initially you only inserted the smaller USB drive and installed AB-S on it. At the time the small USB is /dev/sda. Then you attached a larger USB drive to the router and now it has got precedence during boot-up so it got /dev/sda and the smaller USB drive is now /dev/sdb

I don’t know if there is logic in AB-Solution’s code to handle this properly, maybe @thelonelycoder can chime in here.

 

[thelonelycoder]

I don’t know if there is logic in AB-Solution’s code to handle this properly, maybe @thelonelycoder can chime in here.

AB corrects paths to itself and Entware if it is started with the full path to ab-solution.sh when for some reason the device name or label changed. What other scripts do is outside of AB's realm.

Label yor devices BEFORE installing stuff like Entware, AB or other software on your routers.
It is a well known limitation that when more than one device is plugged in, the name assigned to a device is in order when it's first seen during boot.

 

[deddc23efb]

The leaks seems to be highly dependent on the environment/configuration. I’ve personally not observed any leaks.

And what’s your definition for “safe”? 384.5 fixed a lot of security issues and updated a lot of components, IMHO you should update to be more “safe”.

Hmm interesting. My definition of "safe" is a stable system that doesn't eventually reset or hang because of unknown or not understood bugs and feature interactions. Reading through this thread it seems that 384.5 has these kinds of problem regardless of whatever "security" problems and "updates" were applied. More updates from ASUS does not excite me; I would rather less churn and more ASUS bugs fixed.

I'm glad that you haven't experienced any problems. I'm asking the question because I'm hoping for useful responses from this informed group so that I can make a risk assessment on "upgrading" my own systems.

 

[kfp]

I'm asking the question because I'm hoping for useful responses from this informed group so that I can make a risk assessment on "upgrading" my own systems.

Exactly why I asked you to clarify what “safe” means to you.

Asus writes code like butt and Merlin has been wiping it clean to the best of his ability already. I’d attribute the recent outbreak of problems to Asus adding code to support AiMesh.

Some of us are already testing 384.6 alpha2 and it seems promising.

I mean you can always flash first and just downgrade if it doesn’t work out. If your model is supported you might want to give John’s fork a try since it adds no new functionalities and only backport fixes from stock and Merlin.

 

[Morac]

So I’m still seeing the free memory slowing going down. Down to 290 Mb now. It’s not the httpd server as the memory usage is quite low on that. Is there an easy way to find which process is using the most memory? I tried “top -m”, but that says there are no processes (which is wrong).

 

[kfp]

So I’m still seeing the free memory slowing going down. Down to 290 Mb now. It’s not the httpd server as the memory usage is quite low on that. Is there an easy way to find which process is using the most memory? I tried “top -m”, but that says there are no processes (which is wrong).

Just focusing on the Free memory doesn’t mean much. Instead of top, try ‘free’ which gives you cached memory. If that’s not enough try entware+htop.

 

[ech]

So I’m still seeing the free memory slowing going down. Down to 290 Mb now. It’s not the httpd server as the memory usage is quite low on that. Is there an easy way to find which process is using the most memory? I tried “top -m”, but that says there are no processes (which is wrong).

"grep VmRSS /proc/*/status|sort -k2" will show you the amount of resident memory in use by each process, with the ones consuming most at the end of the output... then you can use "ps|grep $proc" (where $proc is the process number you want to track down, from /proc/$pid/status) to see which process that numeric one is.

 

[ech]

Just focusing on the Free memory doesn’t mean much. Instead of top, try ‘free’ which gives you cached memory. If that’s not enough try entware+htop.

As Morac has said before, the buffers and cached numbers aren't growing, while free memory continues to drop.

 

[Morac]

"grep VmRSS /proc/*/status|sort -k2" will show you the amount of resident memory in use by each process, with the ones consuming most at the end of the output... then you can use "ps|grep $proc" (where $proc is the process number you want to track down, from /proc/$pid/status) to see which process that numeric one is.

Thanks. I’m not seeing anything jump out as being extremely large in size. The top 5 ones are “networkmap” for about 4 MB, three “mastiff” processes at about 3.5 Mb each and httpd at about 3.1 MB.

For completeness free shows, so cached is around 14 MB.:

total used free shared buffers cached
Mem: 515184 215540 299644 0 568 13956
-/+ buffers/cache: 201016 314168
Swap: 0 0 0

Does the networkmap process seem the right size? Also what is “mastiff”?