AC86U Reboot Issue - 384.16

[ApexRon]

Had significant errors on cable modem so I rebooted it to clear counters and re-sync. When it came back online the AC86U said the WAN was disconnected, system log indicating ISP DHCP issue. No worries, I'll just reboot router from GUI and it went through the motions, came back and said WAN was disconnected. Then I noticed my RAM was almost used up when it should have been about 53%. I rebooted the router again and same symptoms. It was then when I noticed the system uptime was in days when it should have been 0.

I resolved by physically powering off the router and then back on. WAN is connected with an IP address assigned, RAM was back to 53%, and uptime was 0.

384.16 must have a bug wherein a router reboot from the GUI no longer reboots router. What it does I don't know.

 

[L&LD]

I've rebooted many RT-AC86U's (many of them remotely over OpenVPN lately) and have not seen that issue.

What scripts, if any, are you running? Do you have any USB devices plugged in? What features and options are you using on the router?

When was the last time the router was fully reset to factory defaults and minimally and manually configured to secure it and connect to your ISP?

 

[ApexRon]

What scripts, if any, are you running?

None

Do you have any USB devices plugged in?

Thumb drive but why would that prevent a reboot? I would expect firmware to unmount as part of reboot or make drive inaccessible after reboot. I will remove and retest.

What features and options are you using on the router?

Nothing I haven’t used for the past six or more firmware versions.

When was the last time the router was fully reset to factory defaults and minimally and manually configured to secure it and connect to your ISP?

Long time but I would expect a power on reset not to work as well.

 

[L&LD]

If memory is so low that a thumb drive is preventing a reboot, I don't know the technical reason, but I have seen it in practice.

Try 'safely removing drive' from the GUI. If that doesn't work, turn off the power by pushing the power button (if your router has one). If that doesn't work, then unplug the power cord.​

Nothing you haven't used for the past six or more firmware versions?

Sorry, my crystal ball is still cloudy. Need actual details.
​

If by 'power-on reset' you mean physically rebooting the router by powering down or unplugging the power plug, yes that works because the RAM content is 'lost' and the router needs to go through the boot-up process to be operational again.

With 'six or more' firmware versions being 'dirty upgrades', I would consider doing an M&M Config and ideally also a Nuclear reset to get your router back to a good/known state.

Please see my signature below for links to those guides and more.

HTH.

 

[ApexRon]

If memory is so low that a thumb drive is preventing a reboot

Last year, when I was running SMB, a GUI reboot worked every time.

'power-on reset'

Power off, turn yourself around three times, power on.

Here's the thing:

1. Router was working just fine. Memory was at 53%
2. Cable modem was rebooted to get it to re-sync to ISP
3. Cable modem came back but router could not connect to WAN. System log said DHCP issue with ISP

Right there that tells me something could be wrong with firmware.

I rebooted the router from the GUI. The screen changed to the 'progress bar' and after a minute or so the router was back with the GUI screen. System up time had not been reset and my free RAM was minimal. To me that says there could be something wrong with the firmware.

Prior to 384.16 I could reboot cable modem every five minutes and the router handled it. Now it cannot. Why?

 

[ColinTaylor]

1. Cable modem came back but router could not connect to WAN. System log said DHCP issue with ISP

Right there that tells me something could be wrong with firmware.

Indeed. But is it reproducible or was it just a one-off?

I rebooted the router from the GUI. The screen changed to the 'progress bar' and after a minute or so the router was back with the GUI screen. System up time had not been reset and my free RAM was minimal. To me that says there could be something wrong with the firmware.

The "progress bar" is just a timer on the web page. It doesn't actually reflect what is happening on the router in any way.

There have been plenty of reports for different models and firmware releases where user's routers inexplicably experienced problems with lack of memory and/or high CPU usage. In most cases (AFAICT) it coincided with high iowait activity. Personally, I have a suspicion it's linked to attempts to download updated AiProtection signatures.

 

[ApexRon]

But is it reproducible or was it just a one-off?

Once I post this reply, I will try to reproduce and update.

The "progress bar" is just a timer on the web page.

I know that but the point is that the router had begun the reboot process far enough to do that.

...linked to attempts to download updated AiProtection signatures.

Not using AiProtection.
**************************************************************
Rebooted cable modem and monitored router. Here is what I saw:

1. Cable unplugged
2. Disconnected
3. Then (this is weird) the WAN IP Address changed to 192.168.100.10 which is not my NAT network nor my cable modem's address, though it is part of my cable modem's network
4. A few seconds after cable modem returned to it's normal GUI, the router WAN address became a 75.xxx.xxx.xxx address which is normal.
5. Rebooted router from GUI
6. Everything came back. All normal

So the answer is, NO I could not recreate and I am not going to attempt to recreate any more. Just reaching out to see if anyone else has experienced this issue. Edited post to add 5 &6.

Here is the router System Log for the erroneous IP address:
Apr 15 12:41:27 nat: apply nat rules (/tmp/nat_rules_eth0_eth0)
Apr 15 12:41:27 wan: finish adding multi routes
Apr 15 12:41:27 dhcp_client: bound 192.168.100.10/255.255.255.0 via 192.168.100.1 for 30 seconds.
Apr 15 12:41:28 WAN_Connection: WAN was restored.
Apr 15 12:43:13 rc_service: udhcpc 1775:notify_rc start_dhcp6c
Apr 15 12:43:13 nat: apply nat rules (/tmp/nat_rules_eth0_eth0)
Apr 15 12:43:13 wan: finish adding multi routes
Apr 15 12:43:13 dhcp_client: bound 75.xxx.xxx.xxx/255.255.240.0 via 75.xxx.xxx.xxx for 45763 seconds.

 

[john9527]

I know that but the point is that the router had begun the reboot process far enough to do that.

The progress bar is put up before the call is made to reboot.

 

[ApexRon]

The progress bar is put up before the call is made to reboot.

Good to know

 

[ColinTaylor]

Then (this is weird) the WAN IP Address changed to 192.168.100.10 which is not my NAT network nor my cable modem's address, though it is part of my cable modem's network

That's perfectly normal for a cable modem. It's just a temporary address given to the client before the connection to the internet is established.