Webinterface crashes

  • ATTENTION! As of November 1, 2020, you are not able to reply to threads 6 months after the thread is opened if there are more than 500 posts in the thread.
    Threads will not be locked, so posts may still be edited by their authors.
    Just start a new thread on the topic to post if you get an error message when trying to reply to a thread.

lbradio

Regular Contributor
I've got this issue with a RT-AC68U, 378.55.

The webinterface crashes, no menu or data is shown.
This happens within an hour after a reboot.
Rebooting the router solves the issue (for a while).

I've already tried disabling OpenVPN, and all TrendMicro stuff, but can't find what causes the crash.
Router can be reached using SSL.
So I was able to read the last entries in the logfile:

Code:
Sep 11 12:03:29 watchdog: restart httpd
Sep 11 12:03:29 rc_service: watchdog 564:notify_rc start_httpd
Sep 11 12:03:29 RT-AC68U: start httpd
Sep 11 12:03:59 watchdog: restart httpd
Sep 11 12:03:59 rc_service: watchdog 564:notify_rc start_httpd
Sep 11 12:03:59 RT-AC68U: start httpd
Sep 11 12:04:09 dropbear[3007]: Child connection from 10.8.1.2:50761
Sep 11 12:04:12 dropbear[3007]: Password auth succeeded for 'admin' from 10.8.1.2:50761
Sep 11 12:04:29 watchdog: restart httpd
Sep 11 12:04:29 rc_service: watchdog 564:notify_rc start_httpd
Sep 11 12:04:29 RT-AC68U: start httpd
Sep 11 12:05:29 watchdog: restart httpd
Sep 11 12:05:29 rc_service: watchdog 564:notify_rc start_httpd
Sep 11 12:05:29 RT-AC68U: start httpd

Any suggestions in how to solve this one?
 

iiaXeii

Occasional Visitor
I have (had) similar issues with my AC87R using the latest Merlin firmware. I have been up for almost 2 days now without issues (so far)... my history

- Was on Merlin latest firmware - was crashing every hours
-Reset to default settings on Merlin (reset was held for 30 sec.) did not help, kept crashing every hours or so.
-Went back to Asus stock latest firmware, was up for a few days
-Went back to Merlin latest firmware BUT:

I did the checksum... which I never did before... so maybe was corrupted? Hard to tell, most function were all up and running... but who knows
I currently only have the PPPoE sessions, I have not enabled anything else... (originally had 2 openvpn clients, bunch of mac addressed reserved)

So far so good... my current thinking is that it may have been corrupted, I know I never did resest in between firmware, the fact that they kinda worked out of the box after flashing didn't lead me in thinking I should since all my settings were still there and working...

so hopefully, it stays this way and next time I update, I will hard reset in between to be sure... but again, after re-flashing Merlin's, it did not crash after an hour like it did before.

There is this thread here about the same thing...:
http://www.snbforums.com/threads/rt-ac68u-often-crash.25477/

Which I posted in commenting that I had the same issues for my AC87R... However, it seems to be stable now or has been for a little while without crashing every hours.
 

lbradio

Regular Contributor
Thank you both for the input. The router is not overclocked, it seems nvram is not full ( 48835 / 65536 bytes)
I have multiple routers (68U / 87U) running the same firmware, this is the only one giving me the headache ;-)

I am pretty sure I did reset to default settings when switching between firmware, but I think I will start over again with this specific router.

I think the procedure was:
- Factory reset by webinterface
- Reboot
- SSH mtd-erase2 nvram
- Power cycle

Correct?

Thanks.
 

lbradio

Regular Contributor
Tonight I erased all settings. Factoryreset, reboot, clear nvram, power cycle.

Afterwards manually reconfigured the router.

How suprised I am the webinterface crashed again after a while.
I'm running out of ideas.

Anyone a good one?

Tx.
 

john9527

Part of the Furniture
I accidentally created the exact error you are seeing when I was doing some work on my fork changing the http port and it couldn't create the listening socket. Any chance of an address conflict with the router address or are you trying to run a web server on http port 80?
 

lbradio

Regular Contributor
I accidentally created the exact error you are seeing when I was doing some work on my fork changing the http port and it couldn't create the listening socket. Any chance of an address conflict with the router address or are you trying to run a web server on http port 80?
Hello John,
Thanks for your comment. I've checked this one, and I am pretty sure no other application is using port 80.

When disconnecting modem the router stays alive.
It seems the modem is messing up the signal in one way or another.
 

lbradio

Regular Contributor
Try changing the port. just for grins.
Do you know if it is possible to change the default webport on the Asus?
Or did I misunderstood you?
 

Calisro

Senior Member
Well, I was thinking to leave it running on 80 and start a new instance of the webconsole on another port. For grins it would be interesting if one stays up while the other (on 80) crashes. I seriously think this is a conflict or something. It sure seems like the console crashes but doesn't free up port 80 correctly and then the daemon tries to restart it and fails. Theory....

logon via ssh
Code:
cd /www
httpd -p 8080 > /jffs/web.log 2>&1 &

If it does crash, however, let's see that web.log file. It will spit out standardout/error to that file.
 

Calisro

Senior Member
What else would be interesting is if you restart the default webconsole from ssh with standard error/logging put to a file and let it crash.

Code:
cd /www
killall httpd
httpd -p 8080 > /jffs/web8080.log 2>&1 &
httpd -p 80 > /jffs/web80.log 2>&1 &

Perhaps this will give us some info on when it crashes and why it can't start.
 

RMerlin

Asuswrt-Merlin dev
Well, I was thinking to leave it running on 80 and start a new instance of the webconsole on another port.

Simply enable both HTTP and HTTPS. The HTTPS instance runs separately from the HTTP one.
 

kvic

Part of the Furniture
Simply enable both HTTP and HTTPS. The HTTPS instance runs separately from the HTTP one.

Merlin, you shall consider porting the feature of configurable HTTP port and not binding to all interfaces from John's fork in V14. It'll be another differentiator in asuswrt-merlin. Your power users will appreciate it.

Personally I'm very happy to get rid of another small script hack of mine once this feature is available in a future release.
 

john9527

Part of the Furniture
consider porting the feature of configurable HTTP port and not binding to all interfaces from John's fork in V14
I'd at least wait a bit....The base change may be a one-liner, but the concept of having it fixed to port 80 was a bit more invasive. I thought I had caught them all, but found a few more which will be in V14E1. And it created what I think is a race condition during factory reset that I needed to workaround.
Also, with the change in logon to the token based logon, things are different in the latest code. The simple things aren't always so simple.

But, at least you know I read some of the threads :)
 

RMerlin

Asuswrt-Merlin dev
Merlin, you shall consider porting the feature of configurable HTTP port and not binding to all interfaces from John's fork in V14. It'll be another differentiator in asuswrt-merlin. Your power users will appreciate it.

Personally I'm very happy to get rid of another small script hack of mine once this feature is available in a future release.

It's not as simple as it seems. Changing the port on the HTTP interface can break things with AiProtection, which is only able to handle non-standard https ports. That's why I held off on changing this for so long.
 

lbradio

Regular Contributor
What else would be interesting is if you restart the default webconsole from ssh with standard error/logging put to a file and let it crash.

Code:
cd /www
killall httpd
httpd -p 8080 > /jffs/web8080.log 2>&1 &
httpd -p 80 > /jffs/web80.log 2>&1 &

Perhaps this will give us some info on when it crashes and why it can't start.

Unfortunately without success ;-)
After the interface crashes, a simple httpd or httpd -p 80 will not show the whole webinterface.
Logonpage is back, but it will only show a part of the main screen.

The log isn't very helpful as well, since it only tells "port already in use". But I assume this is watchdog which already started the interface.
 

Attachments

  • screen 1.JPG
    screen 1.JPG
    37.3 KB · Views: 658

john9527

Part of the Furniture
Just to confirm.....
- gui works and is accessible after a reboot
- sometime (within an hour) the gui is inaccessible and httpd is restarting

So a guess....
Networkmap continues to run in the background loading client names and characteristics. Do you have any special characters in the client names . If so, try changing your client names to be only alphanumeric and - or _ (I see you're using a language enabled setting, so no accented/extended characters either). I've not heard of it crashing httpd, but know it can cause the web page to error and not display. Maybe the sequence of characters is just right to cause a more catastrophic error.
 

tigs

New Around Here
Unfortunately without success ;-)
After the interface crashes, a simple httpd or httpd -p 80 will not show the whole webinterface.
Logonpage is back, but it will only show a part of the main screen.

The log isn't very helpful as well, since it only tells "port already in use". But I assume this is watchdog which already started the interface.
Are you able to access the router through SSH and/or telnet?

I have this similar problem lately. the web UI will all of a sudden not accessible. and so is the SSH. How ever the router is distributing IPs to client and providing internet. A restart of the router will bring back the access. I did not have this issue with previous versions.
 

lbradio

Regular Contributor
Just to confirm.....
- gui works and is accessible after a reboot
- sometime (within an hour) the gui is inaccessible and httpd is restarting

So a guess....
Networkmap continues to run in the background loading client names and characteristics. Do you have any special characters in the client names . If so, try changing your client names to be only alphanumeric and - or _ (I see you're using a language enabled setting, so no accented/extended characters either). I've not heard of it crashing httpd, but know it can cause the web page to error and not display. Maybe the sequence of characters is just right to cause a more catastrophic error.

John, I was pretty sure there were no special characters.
But somehow you pointed me on the devices.
I ended up with disconnecting a HP MFP printer.
Now, for over 24 hours, no problems anymore.

So it seems I've located the issue, but still can't understand why a networkprinter can crash a router...
 

Similar threads

Sign Up For SNBForums Daily Digest

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