Core 2 100% Utilization on 380.61 RT-AC68U

[rlfromm]

The router is doing nothing yet Core 2 is 100% utilized. Any ideas?

Also remote SSH does not work and remote GUI does not work. I have checked all the settings and they are set as they were when these features were working before. Any ideas?

So I noticed the router would no reboot when requested. keep seeing message: Skip event: reboot, in the log.

I had my wife cycle the power and now everything is OK.

Router was up for 11 days. Memory leak of type causing my problem?

 

[RMerlin]

Check with "top" over SSH to determine what is using the CPU.

 

[rlfromm]

I could not check with top. the local SSH or telnet would not work either. I could connect but it would not take any commands.

 

[evh909]

Are you utilizing the dnla server? If so, disable it and see if your CPU usage goes down

 

[avi68]

Im having the exact same problem. I have this problem after 20 days, once the router has been running for 21 days the core 1 is stuck at 100%. I can login to ssh but when I type in top nothing appears.

The same thing happened to me on the previous asuswrt merlin build. When I installed the firmware I did a full factory default reset. Not sure what it is.

Problem is solved after rebooting the router, but then after 20 days then issue appears again.

 

[john9527]

Problem is solved after rebooting the router, but then after 20 days then issue appears again.

Do you have a large ext3/4 formatted USB attached? Check the file system mount count/interval check with tune2fs.

 

[octopus]

WAN - DDNS
Forced refresh interval (in days) 21
Checking every 21 days.

 

[avi68]

Do you have a large ext3/4 formatted USB attached? Check the file system mount count/interval check with tune2fs.

Yes I have a large USB hard drive connected but I think its format is ntfs not ext.

 

[avi68]

WAN - DDNS
Forced refresh interval (in days) 21
Checking every 21 days.

Yes mine is set to 21 days. Is this causing the issue?

If so how do I correct it?

 

[john9527]

Yes mine is set to 21 days. Is this causing the issue?

If so how do I correct it?

The predefined providers I'd say are pretty well debugged. Are you using a custom script? If so, I'd look there for a problem.

@octopus - Great observation!

 

[sfx2000]

The predefined providers I'd say are pretty well debugged. Are you using a custom script? If so, I'd look there for a problem.

@octopus - Great observation!

Something is definitely hung up somewhere is one core is railed, and the remaining core can't respond to interactive inputs...

I used to have a watchdog task that would check certain things, and if the dog were kicked (hung process for example) , it would do two things - grab certain log lines (like tail -n 100 /var/log/syslog) and some stack/mem data, etc - drop that into a location on flash, and then do a sysv init runlevel 6... - and this was turned on by watchdog.sh which was enable|disable (default was disable)...

I'll have to go back and look thru the GPL, but if I recall, there is a watchdog type of task there just for things like this under the hood...

 

[avi68]

The predefined providers I'd say are pretty well debugged. Are you using a custom script? If so, I'd look there for a problem.

@octopus - Great observation!

No I'm not using any custom scripts, just the standard options from the UI

 

[avi68]

Something is definitely hung up somewhere is one core is railed, and the remaining core can't respond to interactive inputs...

I used to have a watchdog task that would check certain things, and if the dog were kicked (hung process for example) , it would do two things - grab certain log lines (like tail -n 100 /var/log/syslog) and some stack/mem data, etc - drop that into a location on flash, and then do a sysv init runlevel 6... - and this was turned on by watchdog.sh which was enable|disable (default was disable)...

I'll have to go back and look thru the GPL, but if I recall, there is a watchdog type of task there just for things like this under the hood...

The only way I think I could test is to change the amount of refresh interval days to like 2 and see if the same problem occurs after 2 days... Just done a full reboot and everything's back to normal.

 

[sfx2000]

The predefined providers I'd say are pretty well debugged. Are you using a custom script? If so, I'd look there for a problem.

Funny how DDNS is popping up the last couple of days - there was another thread where someone was asking to add Cloudflare as a DDNS, and RMerlin had a good response - basically due to goofiness at the DDNS provider level (API changes, etc...) - gist of the response for his build was - nano is there, and that gives folks the chance to write their own script

 

[sfx2000]

The only way I think I could test is to change the amount of refresh interval days to like 2 and see if the same problem occurs after 2 days... Just done a full reboot and everything's back to normal.

That would be a good first step...

 

[john9527]

No I'm not using any custom scripts, just the standard options from the UI

What DDNS provider?

 

[avi68]

What DDNS provider?

No-ip

 

[avi68]

That would be a good first step...

Ok changed the interval to 2 days. Let's see what happens........

 

[Andyf66]

I've seen something very similar on my AC-87U about 3 times so far, although I'm not using DDNS. Only a hard power cycle fixes it, as a soft reset just hangs. Not sure when I first saw it though.

One core locked at 100%. Top reports 50% idle, 50% Nice. Couldn't figure out which process was causing it though. Load average was about 5 or 6. I think it might have been tftpd, but that is a wild guess.

 

[avi68]

I've seen something very similar on my AC-87U about 3 times so far, although I'm not using DDNS. Only a hard power cycle fixes it, as a soft reset just hangs. Not sure when I first saw it though.

One core locked at 100%. Top reports 50% idle, 50% Nice. Couldn't figure out which process was causing it though. Load average was about 5 or 6. I think it might have been tftpd, but that is a wild guess.

What does your router uptime go to before you notice this problem?

At least you can see what processes are running, mine just doesnt respond the any commands in SSH.

My load average also stays at about 4-5 when this problem occurs.