Release [Fork] Asuswrt-Merlin 374 LTS release 49E4

[frazor77]

When I moved to the LTS builds (~9 months ago) I factory reset and haven't done anything outside of the core features... but just in case will SSH in

 

[frazor77]

Command run, no output

 

[john9527]

I was on 48E2 -> 49D6 -> 50D6 (have been updating in the hope something in a newer build fixes it, but its still happening...

Please let me know what data/info I can supply to help diagnose this. syslog attached.

In your syslog I see multiple instances of the httpd server not being available, and the last reboot was caused by the watchdog process going away....very strange.
Since you are running a MIPS based router, when was the last time your reformatted JFFS? (because of how JFFS is allocated on MIPS, as the size of the firmware grows, it can corrupt JFFS).
Only other thought is that you may be suffering from the effects of a power supply that is starting to get weak from age.

 

[frazor77]

I'm not sure when I last reformatted the JFFS.. maybe when I switched to LTS (I would likely have done it if it was specified in the guide to moving from the last Merlin build to the Merlin-LTS builds), but maybe not.

Would it help to save JFFS -> format JFFS on reboot -> restore JFFS? (although I seem to remember it being suggested that doing that is somewhat a 'last resort'?)

I could also see if I have a spare PSU that's compatible... given the router has been in continued 24x7 service for ~8 years, its had a good run!

 

[ColinTaylor]

Would it help to save JFFS -> format JFFS on reboot -> restore JFFS?

Yes, do this. Arguably there's no point in restoring the JFFS files if you're not using any add-on scripts. The only files are likely to be old syslog files. You can look the the backup file with something like 7-Zip to see if there's anything you want restored.

 

[john9527]

Would it help to save JFFS -> format JFFS on reboot -> restore JFFS? (although I seem to remember it being suggested that doing that is somewhat a 'last resort'?)

Did you move any of the 'certs' to jffs to free nvram space? In the syslog I saw a variety of httpd check errors...a timeout and two different cert verification fails. If so, it may be worthwhile to restore those from a different backup if available.

 

[frazor77]

I did not. This is pretty much a stock setup of the latest LTS build.

 

[zsero]

Hi I just upgraded a AC56U from an old release to 49E4. How can I make a full reset? I did a factory restore from GUI + a "Format JFFS on next boot". Is there anything else I should do? I'd just like to start clean.

 

[tsanga]

RT-AC66U here.

Recently, I’ve been experiencing spontaneous reboots every week or so. In syslog, I always see the reboot prefaced by:

watchdog02: no watchdog, restarting

Or other times, not necessarily resulting in a reboot but didn’t notice this before:

watchdog: restart httpd, error detected or process not responding (12)

The last stable release for me was 46Ex, when it could go 2-3 months easily and remain stable. I don’t remember running 47Ex very long, but 48Ex and 49Ex have been giving me this issue. I have not changed any settings, and rolling back to 46E9 restores the stability. I always save jffs, format and restore with each FW update.

Can anyone offer any additional tips for me to troubleshoot this? Is it possible to get more clues from the log?

 

[L&LD]

If you're restoring old backups, don't.

Test with a fresh reset and minimal and manual configuration. And without 'blindly' using old features, scripts or settings that worked 'once upon a time'.

 

[ColinTaylor]

RT-AC66U here.

Recently, I’ve been experiencing spontaneous reboots every week or so. In syslog, I always see the reboot prefaced by:

watchdog02: no watchdog, restarting

Or other times, not necessarily resulting in a reboot but didn’t notice this before:

watchdog: restart httpd, error detected or process not responding (12)

The last stable release for me was 46Ex, when it could go 2-3 months easily and remain stable. I don’t remember running 47Ex very long, but 48Ex and 49Ex have been giving me this issue. I have not changed any settings, and rolling back to 46E9 restores the stability. I always save jffs, format and restore with each FW update.

Can anyone offer any additional tips for me to troubleshoot this? Is it possible to get more clues from the log?

This looks to be exactly the same problem reported by @frazor77 in the posts above yours. Can you upload your syslog to pastebin for us to look at?

 

[keturidu]

This looks to be exactly the same problem reported by @frazor77 in the posts above yours. Can you upload your syslog to pastebin for us to look at?

Got exactly the same problem several months ago. Router keeps restarting (mostly on same 3 a.m.) on it's own. Model: RT-AC66U

sadfjwx - Pastebin.com

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

pastebin.com

 

[ColinTaylor]

Got exactly the same problem several months ago. Router keeps restarting (mostly on same 3 a.m.) on it's own. Model: RT-AC66U

sadfjwx - Pastebin.com

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

pastebin.com

Thanks for that. FYI it's not restarting at 3 am, that's just the default date/time the router gives itself when it boots up. The reboot times in your log file were:

Jul  6 04:51:23 watchdog02: no watchdog, restarting
Jul  9 11:36:34 watchdog02: no watchdog, restarting
Jul 12 15:11:45 watchdog02: no watchdog, restarting

 

[john9527]

Got exactly the same problem several months ago. Router keeps restarting

For your case, I'd definitely start with a jffs reformat....

Jul  5 00:54:06 watchdog: restart httpd, error detected or process not responding (12)
Jul  5 00:54:06 rc_service: watchdog 387:notify_rc stop_httpd
Jul  5 00:54:06 rc_service: watchdog 387:notify_rc start_httpd
Jul  5 00:54:06 rc_service: waiting "stop_httpd" via watchdog ...
Jul  5 00:54:07 httpd: start httpd
Jul  5 00:54:07 kernel: jffs2_flush_wbuf(): Write failed with -5
Jul  5 00:54:07 kernel: Write of 622 bytes at 0x00da4680 failed. returned -5, retlen 0
Jul  5 00:54:07 kernel: Not marking the space at 0x00da4680 as dirty because the flash driver returned retlen zero

 

[tsanga]

This looks to be exactly the same problem reported by @frazor77 in the posts above yours. Can you upload your syslog to pastebin for us to look at?

Here you go. Thank you for taking a look.

RT-AC66U log 3.0.0.4.374.43_49E4j9527 - Pastebin.com

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

pastebin.com

In my case, syslog captured the beginning of one spontaneous reboot, then there was another one on Jul 9.

 

[zsero]

Hi I just upgraded a AC56U from an old release to 49E4. How can I make a full reset? I did a factory restore from GUI + a "Format JFFS on next boot". Is there anything else I should do? I'd just like to start clean.

So I'm a bit confused about factory resets. I tried pressing the reset pin on the back and nothing happened. The WPS button reset worked.
Then I upgraded firmware and did restore factory defaults. Then I did JFFS format on next boot.
What does each of these do? Why did the reset button not do anything, while the WPS button worked?

 

[Rahul_K]

Here you go. Thank you for taking a look.
EDIT: [LINK REMOVED]
In my case, syslog captured the beginning of one spontaneous reboot, then there was another one on Jul 9.

Count me in among those whose RT-AC66U running 3.0.0.4.374.43_49E4j9527 restarts every 56-60 hours due to "watchdog02: no watchdog, restarting". My router logs can be found https://pastebin.com/HmPceqbS . It's worth noting that when the router restarts, the 2 processes wanduck and watchdog both start off with memory consumption looking like this (using top-m):

PID^^^VSZ^VSZRW   RSS (SHR) DIRTY (SHR) STACK COMMAND
  361  3468  1072  1340   516   788     0    84 /sbin/wanduck
  392  3464  1068  1332   532   776     0    84 watchdog

but just before the router restarts both wanduck and watchdog were consuming a a lot more memory. I couldn't get the memory consumption of watchdog as that process had terminated, but that too had ~37k under VSZ after about 56 hours of uptime.

PID^^^VSZ^VSZRW   RSS (SHR) DIRTY (SHR) STACK COMMAND
  361 37288 34892 35160   516 34608     0    84 /sbin/wanduck

While there are plenty of other threads about how linux performs memory-management, dare I say this does feel like a memory leak. I was running my router with the bare minimum configuration (just WAN credentials, WIFI SSID and admin password after a reset) a few weeks back but I got the same behaviour. Now I've set my router to restart every 48 hours.

 

[keturidu]

For your case, I'd definitely start with a jffs reformat....

Jul  5 00:54:06 watchdog: restart httpd, error detected or process not responding (12)
Jul  5 00:54:06 rc_service: watchdog 387:notify_rc stop_httpd
Jul  5 00:54:06 rc_service: watchdog 387:notify_rc start_httpd
Jul  5 00:54:06 rc_service: waiting "stop_httpd" via watchdog ...
Jul  5 00:54:07 httpd: start httpd
Jul  5 00:54:07 kernel: jffs2_flush_wbuf(): Write failed with -5
Jul  5 00:54:07 kernel: Write of 622 bytes at 0x00da4680 failed. returned -5, retlen 0
Jul  5 00:54:07 kernel: Not marking the space at 0x00da4680 as dirty because the flash driver returned retlen zero

[blush] Did that, and after 36 hours - no reboot. Will wait for a few days, just to be sure it won't restart after 56-60 hours, like Rahul_K mentions. Thanks.

 

[frazor77]

So it all seemed OK until this morning... when it rebooted again (fortunately while everyone was asleep). Happy to upload the latest logs if useful? In the mean time also going to see if I have another PSU that will work.

 

[frazor77]

19v @ 1.58A... OK didn't realise it was such an unusual PSU....