RT-AC68U - 100% CPU & Memory use. "mtdblock3" and "tdts_rule_agent"?

[BitingChaos]

I'm thinking this started when the 384 branch came out. Or maybe it was before that. I try to stay up date. It's been doing this for months, and I'd like to get it fixed.

Anyway, whenever I'd notice the network slowing down, I'd try to log into the web interface to check the QoS /traffic page to see if someone was using all the bandwidth.

Many times the web interface would be really slow to load. Both CPU cores would be at 100%. There'd also be almost no free memory. Sometimes the web interface would just stop responding.
Many times SSH wouldn't even connect. It would just time out. Even if SSH connected, the "reboot" command would do nothing. After a few minutes of "nothing" happening, all web connections would then get dropped.

I'd manually press the power button, count to 3, then turn it back on.

After doing this way too many times, I configured the Automatic Reboot feature. Every Sunday night it would restart.
But then the connection seemed to bog down during the end of the week (Friday). To better fit our usage patterns, I moved the reboot to Wednesday night.

Well, then the connection started to bog down on Tuesday. It looks like I might need to move to a daily reboot.

On one of the most recent times I noticed performance issues, I quickly SSH'd in and ran top. Two processes kept alternating for the top-spot of eating all CPU cycles.

tdts_rule_agent -g -f /jffs/signature/rule.trf
[mtdblock3]

One would be taking all processing power, then the other. Eventually the router would stop responding.

I have 384.2_7 installed right now.

I've done both "Restore" and "Initialize" under Factory default. AiCloud stuff is all turned off.

I've tried to change minimal things, to maybe narrow down the issue.

After a factory wipe, I set:

* WiFi on 2.4GHz and 5GHz (different SSIDs)
* enable QoS App analysis and Adaptive QoS

I haven't set up port forwarding, VPN, changed anything with the firewall, etc.

What's the best way to troubleshoot this? Should I just wipe and go back to 374?

 

[BitingChaos]

I'm thinking this started when the 384 branch came out. Or maybe it was before that. I try to stay up date. It's been doing this for months, and I'd like to get it fixed.

I guess this is still an issue.

I've been using the automatic reboot function of the firmware. It cannot stay up longer than a few weeks without all resources completely exhausted.

"mtdblock3" shows up at the top for CPU usage, but only around ~15%.
The web interface showed Core1 and Core2 at 100%, and Mem at 92% before it became too slow to use.

I had factory reset and purchased a beefier power supply, just in case that was the issue.

I haven't been able to find a solution.

 

[Ronald Schwerer]

Run "cat /proc/cpuinfo" when this is happening. What's your BogoMIPs values?

 

[BitingChaos]

CPU0: 1595.80
CPU1: 1599.07

What should they be?

 

[Ronald Schwerer]

CPU0: 1595.80
CPU1: 1599.07

What should they be?

Was that when you saw CPU @ 100%?
That's the expected values for a 68U running w/clkfreq=800,666.
Are you using a VPN or other apps beyond stock?

 

[northumberland]

I've seen similar on AC88U running current firmware, also saw it a couple of revisions ago.

Only appeared to affect 1 core though, but was 100%

Running top showed openvpn at the top of the list, but only using a small amount but if the CPU. No one was connected to the VPN at the time.

It was almost like the router was getting confused that the CPU was slower that it really was, thus 100% was shown when only 5% was in use. (I hope that makes sense)

Only non stock stuff running is skynet.

Only happened once on current firmware but thought I'd mention it here in case it's related.

 

[joesixpack]

Just ran into this issue. Only starting showing up after installing AdGuard Home a few days ago. On latest Merlin.