What's new

Routers need hard reboot (powercycle) as they hang

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

KPUK

Occasional Visitor
I've had this with all versions of the Merlinf/w, right up to the latest and with both AC66U_B1 and AC88U routers. Every few weeks, they need a hard power cycle in order to be able to log on to the web interface OR for their DHCP server to work. Looks like some services must just hang or stop and as I can't access them remotely, I'm not able to check anything out. They continue to route traffic, just not service newly connected devices with IP addresses or renew existing devices whose IPs have expired and don't let me in to the web interface.

I'd kind of hoped/thought this was an issue with a specific model (say the 66U_B1) or a particular firmware version, but as this has been the case now for a couple of years and cross-device and over many f/w upgrades, it's obviously not.

Anyone any thoughts? It's kind of a bit annoying. I did put in scheduled reboots just to clear things out, but once or twice I've been unlucky in that they still hang even then. I also don't like scheduling reboots as a solution to something that's obviously related to a software issue.

Thanks.
 
First thing I'd be checking is for any signs of a memory leak (i.e., available memory dropping noticeably and consistently over time).
 
I don't see this as a general issue for many RT-AC66U_B1's and RT-AC3100's (sister model to the RT-AC88U).

When was the last time the router(s) were fully reset and manually setup? With no backup config files restored or old settings 'blindly' input because they 'once worked' either.

You may want to see the last two paragraphs of the link below to get your routers back to a good/known state (they are not there now).

 
I have the same problem. Had to hard reboot my AC 88U every 10 days or so. I am happy to collect any info u needed to troubleshoot this .
 
I just putty in and "service restart_httpd" happens frequently if Im messing with the gui alot.
 
AC88U similar issue - needs a hard reset now and then when my NAS becomes disconnected from the router. Not sure what is causing it, although the last time this happened was when the virtualisation app on the NAS was updated. The NAS is lan port connected. Everything else seems fine, just the NAS can't reconnect. S/w rebooting via the GUI doesn't work - need to power off and on again. Updating f/m doesn't work either - had the same issue on newly installed f/m (f/m fully reset via WPS switch and numerous reboots, reformats of partitions etc...)
Rebooting NAS doesn't work, so must be router side.
 
Problem exists for me as well. Usually every 3-4 weeks or so. Weird, but a reboot fixes. In my case, F/W reboot does work though.

Memory isn't leaking out as far as I can tell, it used to with the TrendMicro AI enabled.
 
The only thing I can think of in my case is something (whatever that is) is becoming choked up in the hardware and only powering off the router can clear it. E.g. buffers at the LAN interface port.
If the issues do not show up in the GUI and therefore f/w, such as a memory leak or something else, then the problem would seem to be at the h/w level?

@SuperDuke I do have QOS / app analysis enabled but only to monitor bandwidth usage - I'll disable this and see if the problem comes back
 
Last edited:
Since 384.14 onwards I've had issues with both an AC88U then an AX88U where all of a sudden a device I go to use has "lost internet" - when having a look at the device it can't get an IP address from DHCP. Have you tried looking in the logs? Commonly in my logs I can see the device asking for an IP but it never gets offered one - so multiple dnsmasq DHCP discovers and dnsmasq DHCP requests but no dnsmasq DHCP offer, sometimes pages of it. Sometimes the GUI is OK, sometimes it is inaccesible, reboot sorts it. Sometimes it can happen in a few days, other times weeks or even months.

Yes I have tried full resets both reset buttton or WPS reset and manual basic config. Yes very hard to fault find on something that can take weeks or months to happen again.

I recently updated to 384.19 with a full reset - had a few issues with WiFi stability initailly and a problem with Quad9 DoT DNS (had worked for months without issue!!) but think I did have the DHCP issue once (hard to tell with the WiFi and DNS issues at the same time, but (hugely tempting fate) it's been up for 16 days now without issue....... time will tell!
 
First thing I'd be checking is for any signs of a memory leak (i.e., available memory dropping noticeably and consistently over time).

Yeah - bit of a pain, though, if I have to keep logging in to check memory consumption. I guess that was my question - are there any known issues such as memory leaks that makes this firmware unreliable? I can guarantee that my routers will lock up if I just leave them alone and that's not exactly what you want from a router...
 
I don't see this as a general issue for many RT-AC66U_B1's and RT-AC3100's (sister model to the RT-AC88U).

When was the last time the router(s) were fully reset and manually setup? With no backup config files restored or old settings 'blindly' input because they 'once worked' either.

You may want to see the last two paragraphs of the link below to get your routers back to a good/known state (they are not there now).


Fully resetting and manually re-configuring isn't really something I want to do as there are a lot of rules, etc.

In something like a router I just want to be able to leave it running and rely on it. Having to "tend to it" regularly basically makes it too unreliable for me.

If you (or others) aren't seeing the same issues, I would imagine it's because we're making use of different modules. For example, I have a PPTP server running. You may not, etc., etc.
 
The only thing I can think of in my case is something (whatever that is) is becoming choked up in the hardware and only powering off the router can clear it. E.g. buffers at the LAN interface port.
If the issues do not show up in the GUI and therefore f/w, such as a memory leak or something else, then the problem would seem to be at the h/w level?

@SuperDuke I do have QOS / app analysis enabled but only to monitor bandwidth usage - I'll disable this and see if the problem comes back

When it locks, it continues to route, but I can't even access the GUI to check things out. A power cycle is the only option I have.
 
Since 384.14 onwards I've had issues with both an AC88U then an AX88U where all of a sudden a device I go to use has "lost internet" - when having a look at the device it can't get an IP address from DHCP. Have you tried looking in the logs? Commonly in my logs I can see the device asking for an IP but it never gets offered one - so multiple dnsmasq DHCP discovers and dnsmasq DHCP requests but no dnsmasq DHCP offer, sometimes pages of it. Sometimes the GUI is OK, sometimes it is inaccesible, reboot sorts it. Sometimes it can happen in a few days, other times weeks or even months.

Yes I have tried full resets both reset buttton or WPS reset and manual basic config. Yes very hard to fault find on something that can take weeks or months to happen again.

I recently updated to 384.19 with a full reset - had a few issues with WiFi stability initailly and a problem with Quad9 DoT DNS (had worked for months without issue!!) but think I did have the DHCP issue once (hard to tell with the WiFi and DNS issues at the same time, but (hugely tempting fate) it's been up for 16 days now without issue....... time will tell!

Thanks. Yes - I've also had issues with DHCP, so mine could be similar...
 
So clearly there are some real issues with the firmware and these issues are not new ones. It's been the case for years. A very big shame because it makes the devices unreliable, which is the last thing you want in a router!
 
I may try that. You using telnet of ssh? Where's the setting to enable one/both of those?
under administration. I use ssh (putty for windows).
 
I've had this problem for a long time with several different Merlin releases. But since I added a cron job and script to run every few hours that does the "service restart_httpd" command, I have not had the problem. Attached is my script. In an SSH session, put the file into /jffs/scripts folder, and set your permissions to 755. Then manually run it to load the cron job by doing a "sh restart-GUI" command. After that, you can verify it is scheduled to run by doing a "cru l" command. After a few hours, look in the logs in the GUI to verify that the file executed every 4 hrs. The script also checks if its included in the services-start file, and if not, adds it. That will take care of loading it after a reboot.

Edit: I had to name the file with a .txt extension in order to load it into this message. For you to use, after you copy it to the scripts folder, rename it by deleting the extension.
 

Attachments

  • restart-GUI.txt
    529 bytes · Views: 160
One thing I did notice, by ssh-ing into the router. I saw hundreds of processes, all with this:

cp /tmp/syslog.log /tmp/syslog.log-1 /jffs

I'm not suggesting that's necessarily THE problem, but it's certainly A problem.
 
Fully resetting and manually re-configuring isn't really something I want to do as there are a lot of rules, etc
Unfortunately, although inconvenient, it may be the best option. It would set you back to a default, clean state. There are very well known issues that *can* be a result of upgrading firmware without a clean reset and reconfigure now and then. These are not enterprise level devices and corruption can occur in the configuration that just gets carried along and eventually causes problems. If you run add-on scripts, or have usb devices on your router, your chances of this happening is even greater. As an example, look at the jffs issues Asus introduced recently.

After reading all the problems people were having with 384.19, I personally am staying with 384.18 until the 386 stream is available in a stable release.

Until you at least try a reset and reconfigure, you will never know if you are looking at a firmware issue, a hardware issue or a simple configuration issue. Knowing that your symptoms are typical of needing to rebuild the configuration may help you make the decision to give it a try :) (and with jffs issues showing in the log, formating the jffs partition as part of the reset wouldn't be a bad thing either)

Get a simple basic setup up and running to provide stability, then add the extra features you want slowly over time to see if one of them causes you issues.
 
I've had this problem for a long time with several different Merlin releases. But since I added a cron job and script to run every few hours that does the "service restart_httpd" command, I have not had the problem. Attached is my script. In an SSH session, put the file into /jffs/scripts folder, and set your permissions to 755. Then manually run it to load the cron job by doing a "sh restart-GUI" command. After that, you can verify it is scheduled to run by doing a "cru l" command. After a few hours, look in the logs in the GUI to verify that the file executed every 4 hrs. The script also checks if its included in the services-start file, and if not, adds it. That will take care of loading it after a reboot.

Edit: I had to name the file with a .txt extension in order to load it into this message. For you to use, after you copy it to the scripts folder, rename it by deleting the extension.

Thanks! I've had this issue on my AC-R3200 running merlin. I've had this issue now for 4-5 years and just set up this script. Fingers crossed!
 

Latest threads

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top