Is my RT-AX86U dying?

[TheMegaMan]

My RT-AX86U seems like it might be dying. I'm on 388.2_2 merlin firmware, with a few scripts added.

I've had a few problems recently that may or may not be related. I'll jot them down here...

- A week or two ago, I saw error entries in the log for 'command failed', seemingly when trying to write to NVM. I assume this is the jffs partition (rather than the SSD i have hanging on the USB port). These seem to have stopped now, though.

- An ongoing problem we've had for some time is that my GF's iPad keeps on disconnecting and reconnecting. It seems like this might be some incompatibilities between iPads and the Asus WiFi6 implementation.

- Router RAM usage seems to be around 99%. I know RAM is there to be used, but being that full without a Swap file doesn't sound like the ideal scenario. I have AMTM installed with Entware, and I tried adding a swap file, and while it seemed to work, it immediately said no swap file existed. So I can't add one.

- The router has started rebooting itself in the last couple of weeks. Actually, it's restarting, but doesn't seem to be a full reset, as it comes back again pretty quickly, as if the radio didn’t restart, just the OS. This can happen as soon as after 15 minutes, or a few hours. I thought it was down to an app on my tablet trying to talk to a Chinese server that the router was blocking (I get an error about redirecting to an 'unavailable' page in the logs roughly when it happens, but it's also restarted when the tablet hasn't been used), but that might be coincidence.

I haven't done so yet, but I'm about to do a factory reset, and set it up clean.

Will this also wipe the jffs partition? If so, I assume I'll need to set up Entware and AMTM again from scratch...which will probably be a good thing.

Is there a way to check the health of the jffs partition before reinstalling any scripts there?

Anyone seen anything like this before and can suggest any tests I can perform to determine whether I have a hardware fault looming, or it's purely software configuration?

I'm contemplating replacing the router with something non-Asus (I'm getting paranoid about the quality of their hardware having had 2 previous routers fail), but I do like Merlin!

Cheers,
Adam

 

[ColinTaylor]

It's possible that your problems are caused by errant addon scripts or a faulty USB drive. Without seeing the log files and doing more diagnosis it's impossible to say.

Doing a hard reset on the RT-AX86U will wipe out the contents of the /jffs partition. I suggest you remove the USB drive and do the hard reset. Then configure the router but do not plug in the USB drive or install any addon scripts. Run for a few days (or as long as necessary) like this to see if you have a stable network.

 

[jtp10181]

The 99% RAM use is a concern, what Colin said is solid advice.

Also, if your USB is a 2.0 drive I would suggest picking up a 3.0 before setting that back up (and use the 3.0 slot on the router). I picked up this one earlier this year and it works much better than my old 2.0 drive I was using: https://www.amazon.com/gp/product/B07D7P4SY4/?tag=snbforums-20

 

[muffintastic]

If you're going to run an external device I suggest buying a NVME drive and a NVME USB 3.0 caddy and plugging it the USB 3.0 port.

Ive noticed in the past when using USB keys it would bog down my router when it was on it's way out since switching I've never had an issue.

 

[TheMegaMan]

It's possible that your problems are caused by errant addon scripts or a faulty USB drive. Without seeing the log files and doing more diagnosis it's impossible to say.

Doing a hard reset on the RT-AX86U will wipe out the contents of the /jffs partition. I suggest you remove the USB drive and do the hard reset. Then configure the router but do not plug in the USB drive or install any addon scripts. Run for a few days (or as long as necessary) like this to see if you have a stable network.

Yeah, that was my thinking, although I'd not changed/updated the scripts for some time before the problems started...hence deducing it might be deteriorating hardware causing the issues. I'd already removed Diversion, and stat gathering scripts, with no improvement.

I've done the hard reset and just minimal configuration to get the router up and running again now, and it's so far lasted over 24 hours without restarting. So that's obviously a good sign. I'll finish the setup tomorrow (adding DHCP manual IP addresses, etc), and let it run for another day before looking at enabling the scripts again.

I have made an observation about the USB storage, which I now suspect could be the root. I'll detail that in my reply to @muffintastic...

Cheers for the advice.

 

[TheMegaMan]

The 99% RAM use is a concern, what Colin said is solid advice.

Also, if your USB is a 2.0 drive I would suggest picking up a 3.0 before setting that back up (and use the 3.0 slot on the router). I picked up this one earlier this year and it works much better than my old 2.0 drive I was using: https://www.amazon.com/gp/product/B07D7P4SY4/?tag=snbforums-20

I can certainly believe that 99% memory usage with no swapfile, could cause unpredictable behaviour when it needs another couple of MB... Since doing the hard reset, it's not sitting at just over 50%, so much more comfortable. I'll keep a close eye on this as I add my desired scripts back again.

I actually had a spare 1TB 2.5" SSD in a USB 3 caddy plugged in, but I had the port configured as USB 2, as I seemed to be suffering terrible data speeds on the 2.4GHz band, and I'd seem a comment that USB 3 could cause this. The problem went away but I'd not yet got around to reverting to USB 3 to prove that was the causse or whether it was some other interference.

 

[TheMegaMan]

If you're going to run an external device I suggest buying a NVME drive and a NVME USB 3.0 caddy and plugging it the USB 3.0 port.

Ive noticed in the past when using USB keys it would bog down my router when it was on it's way out since switching I've never had an issue.

This is interesting. As I just mentioned to @jtp10181. I was actually using an old 1TB 2.5" SATA SSD in a USB 3.0 caddy, formatted to ext3. Speed is not a major issue, so I don't really *need* NVMe (and I suspect USB 3.0 would be the bottleneck, anyway).

However, as part of the reset, I've disconnected the SSD from the router, and plugged it into my PC (with ext2fsd), to back it up. That worked well enough, but I then tried to remove the ext3 partition to reformat it as NTFS and run some tests on it, and it crashed the PC. Now, whenever I plug it into the PC, I get a BSOD. So it's possible the SSD is trash.

Which reminded me that I used a normal USB flash key in another Asus router a while ago, and that died, too.

Which raises the question about whether the Asus software is doing something nasty to the attached USB device that's thrashing it with writes and exceeding the write cycle lifetime.

Curious. But yes, I agree that proper SSD storage should be more robust to this kind of abuse than a cheap USB key, so I'll either dig out another 2.5" SSD or try the NVMe solution.

Cheers!

 

[muffintastic]

This is interesting. As I just mentioned to @jtp10181. I was actually using an old 1TB 2.5" SATA SSD in a USB 3.0 caddy, formatted to ext3. Speed is not a major issue, so I don't really *need* NVMe (and I suspect USB 3.0 would be the bottleneck, anyway).

However, as part of the reset, I've disconnected the SSD from the router, and plugged it into my PC (with ext2fsd), to back it up. That worked well enough, but I then tried to remove the ext3 partition to reformat it as NTFS and run some tests on it, and it crashed the PC. Now, whenever I plug it into the PC, I get a BSOD. So it's possible the SSD is trash.

Which reminded me that I used a normal USB flash key in another Asus router a while ago, and that died, too.

Which raises the question about whether the Asus software is doing something nasty to the attached USB device that's thrashing it with writes and exceeding the write cycle lifetime.

Curious. But yes, I agree that proper SSD storage should be more robust to this kind of abuse than a cheap USB key, so I'll either dig out another 2.5" SSD or try the NVMe solution.

Cheers!

NVME's aren't that expensive these days anyway. As stated, I found that this is a better solution for myself and doesn't create a bottleneck. When I've created a 2 GiB swap file it creates it pretty quick. file transfers are quick too. Both of my routers have been fine since using this method.

 

[Mogsy]

I can certainly believe that 99% memory usage with no swapfile, could cause unpredictable behaviour when it needs another couple of MB... Since doing the hard reset, it's not sitting at just over 50%, so much more comfortable. I'll keep a close eye on this as I add my desired scripts back again.

I actually had a spare 1TB 2.5" SSD in a USB 3 caddy plugged in, but I had the port configured as USB 2, as I seemed to be suffering terrible data speeds on the 2.4GHz band, and I'd seem a comment that USB 3 could cause this. The problem went away but I'd not yet got around to reverting to USB 3 to prove that was the causse or whether it was some other interference.

Always go for minimal config and what you need only. I haven’t made the jump from USB to SSD yet. My ram usage are at 76% now, 89% highest.