Asuswrt-Merlin 380.59 is now available

[jeff288]

It was already at 30 minutes. Here's the recent system log (may there is something here):

May 16 12:49:12 login[32353]: root login on 'pts/0'
May 16 12:49:24 login[32359]: root login on 'pts/0'
May 16 12:49:36 login[32368]: root login on 'pts/0'
May 16 12:49:48 login[32374]: root login on 'pts/0'
May 16 12:50:00 login[32381]: root login on 'pts/0'
May 16 12:50:02 rc_service: httpds 31991:notify_rc restart_time;restart_httpd;restart_upnp
May 16 12:50:02 dropbear[424]: Early exit: Terminated by signal
May 16 12:50:02 kernel: klogd: exiting
May 16 12:50:02 syslogd exiting
May 16 12:50:02 syslogd started: BusyBox v1.20.2
May 16 12:50:02 kernel: klogd started: BusyBox v1.20.2 (2016-05-10 11:44:43 EDT)
May 16 12:50:02 dropbear[32395]: Running in background
May 16 12:50:02 start_nat_rules: apply the nat_rules(/tmp/nat_rules_eth0_eth0)!
May 16 12:50:02 hour monitor: daemon is starting
May 16 12:50:02 RT-N66U: start httpd - SSL
May 16 12:50:02 RT-N66U: start httpd
May 16 12:50:02 miniupnpd[32417]: HTTP listening on port 46272
May 16 12:50:02 miniupnpd[32417]: Listening for NAT-PMP/PCP traffic on port 5351
May 16 12:50:12 login[32418]: root login on 'pts/0'
May 16 12:50:24 login[32424]: root login on 'pts/0'
May 16 12:50:36 login[32433]: root login on 'pts/0'
May 16 12:50:48 login[32439]: root login on 'pts/0'
May 16 12:51:00 watchdog: restart httpd
May 16 12:51:00 rc_service: watchdog 450:notify_rc stop_httpd
May 16 12:51:00 login[32456]: root login on 'pts/0'
May 16 12:51:03 rc_service: watchdog 450:notify_rc start_httpd
May 16 12:51:03 RT-N66U: start httpd - SSL
May 16 12:51:03 RT-N66U: start httpd
May 16 12:51:12 login[32464]: root login on 'pts/0'
May 16 12:51:24 login[32469]: root login on 'pts/0'
May 16 12:51:36 login[32475]: root login on 'pts/0'

Are you using Firefox? I had an issue like this before and it also helped someone else. IIRC, I had to delete places.sql in my FF profile. I'd try a different browser or profile also.

 

[RMerlin]

May 16 12:51:00 watchdog: restart httpd

The web server is crashing, no idea why (Asus has been stuffing so much code into it that it's no longer "just a web server".)

Try accessing it over HTTP instead of HTTPS - there's been some issues with SSL in the latest GPL release, and I'm unable to track them down so far.

 

[RMerlin]

If it helps, a few posts back I could reproduce the httpd restart consistently in the dhcp page across different browsers.

I looked at that particular code path, and this page does nothing special - it's mostly just nvram_get() calls. The most likely culprit is probably tied to the networkmap code. That's been problematic for a while already. Every time I fix one bug in it (and Asus also fixes others on their end), a new one seems to appear.

If for a while they would shift their efforts to port your code into theirs (it would be an admirable humble move), both you and them would probably take advantage of each other in the long run.

Corporate goals don't always match mine, so a lot of my code wouldn't make sense in the stock firmware (for instance, all the custom script support). They also tend to pre-allocate developer resources well ahead of time. Last year I sent them a list of often-requested features that I compiled based on what users were asking most often. They got back to me, a lot of those items were added to their todo list (or already on it), however some of these won't be worked on in the near future as developer resources were already allocated to other projects.

Asus most likely have to plan development a year or so in advance, to match with their hardware releases. They probably keep some wiggle room in there for bug fixes, security fixes, etc... That probably leaves little room for any extra work.

 

[bayern1975]

very strange....with 380.57 firmware i have no pppd connection terminate errors.....i think there is a fault with this asus beta 2697 code.....

 

[kevinnaronha]

The web server is crashing, no idea why (Asus has been stuffing so much code into it that it's no longer "just a web server".)

Try accessing it over HTTP instead of HTTPS - there's been some issues with SSL in the latest GPL release, and I'm unable to track them down so far.

I am using HTTP and it is still crashing.

 

[RMerlin]

I am using HTTP and it is still crashing.

Then something else is causing it to crash, and I don't know what. One thing you can try is to see if you can track it to a specific device on your network that causes the server to crash when that particular device is connected.

 

[cowst]

I looked at that particular code path, and this page does nothing special - it's mostly just nvram_get() calls. The most likely culprit is probably tied to the networkmap code. That's been problematic for a while already. Every time I fix one bug in it (and Asus also fixes others on their end), a new one seems to appear.

Just one last piece of info, in case it helps you deciding in favor or against your current gut feeling about the culprit
I checked with dev tools what happens with the page code itself when it crashes, and I noticed that it all starts with a few icons trying to load:
jquery.js:5 GET https://192.168.2.1:8443/ajax_uploadicon.asp?clientmac=001DEC05B5F1&_=1463429248986 send @ jquery.js:5x.extend.ajax @ jquery.js:5getUploadIcon @ client_function.js:776showdhcp_staticlist @ Advanced_DHCP_Content.asp:259(anonymous function) @ VM1484:1

After 3 of these I get connection refused, which most likely means httpd restarted.
I don't know if the error above happens to cause the restart, or it appears due to the restart.
However I can see no icon for my 5 static devices.

 

[Cwelle]

Hi
Thank you Merlin for all your efforts on this great firmware.

I just updated to 380.59 on AC87U to find all my DLNA devices disappeared. Checked with smart TV, laptops and smart phones all showing nothing. DLNA I miss is synology NAS media server but equally effects other DLNA devices which usually appear (though I don't in practice use) e.g. windows DLNA, logitech server etc. Other networking functions remain functioning as normal.

Downgrading back to previous merlin release on the AC87U fixes the problem and all DLNA devices show up again.

Is there a setting which I am overlooking in the 380.59 firmware or is this a bug?

Many thanks for your help!

I've noticed the same bug in 380.59 beta2. I'm using the same router: AC87U.

 

[RMerlin]

Just one last piece of info, in case it helps you deciding in favor or against your current gut feeling about the culprit
I checked with dev tools what happens with the page code itself when it crashes, and I noticed that it all starts with a few icons trying to load:
jquery.js:5 GET https://192.168.2.1:8443/ajax_uploadicon.asp?clientmac=001DEC05B5F1&_=1463429248986 send @ jquery.js:5x.extend.ajax @ jquery.js:5getUploadIcon @ client_function.js:776showdhcp_staticlist @ Advanced_DHCP_Content.asp:259(anonymous function) @ VM1484:1

After 3 of these I get connection refused, which most likely means httpd restarted.
I don't know if the error above happens to cause the restart, or it appears due to the restart.
However I can see no icon for my 5 static devices.

Could you provide me with the list of icons that are being loaded, in case a specific icon might be the cause?

Did you customize any of the icons, or are they all the default ones provided by networkmap?

 

[RMerlin]

I've moved the general development discussions to a separate thread.

 

[agilani]

rt-ac68p dirty flash/upgrade from 380.58 and its been solid for the past 3 days. It looks like it's fixed some of the cpu issues

 

[unsynaps]

The original issue was that you would get bogus results no matter how NAT acceleration was configured. This issue has been fixed, but that does not change the fact that NAT acceleration must be kept disabled, as it's written on the webui.
.

Ahh thank you. Will give this a shot in the next day or so.

 

[dfran1]

This .59 final has been great! With my ac5300, I turned off smart connect and went to the Highest channels on the 3 bands and I get the best wifi results. IOS devices switch automatically, my main concern is my IOS devices, but my android well does not like to switch bands, though I don't care if it does not.
I use adaptive QOS for gaming mainly. (I don't know if it even does anything, but I seem to get no lag)
This build seems outstanding for me.
Thank You Merlin!!

 

[stevef9432203]

I believe there is a problem with Media Server.
Network devices (smartphones and SmartTVs) NOT see the Media Server.
I did several tests and concluded that the problem is in this new firmware.

I can confirm beta2 and release of 380.59 don't show any of the content on the usb drive,.

 

[RMerlin]

Just one last piece of info, in case it helps you deciding in favor or against your current gut feeling about the culprit
I checked with dev tools what happens with the page code itself when it crashes, and I noticed that it all starts with a few icons trying to load:
jquery.js:5 GET https://192.168.2.1:8443/ajax_uploadicon.asp?clientmac=001DEC05B5F1&_=1463429248986 send @ jquery.js:5x.extend.ajax @ jquery.js:5getUploadIcon @ client_function.js:776showdhcp_staticlist @ Advanced_DHCP_Content.asp:259(anonymous function) @ VM1484:1

After 3 of these I get connection refused, which most likely means httpd restarted.
I don't know if the error above happens to cause the restart, or it appears due to the restart.
However I can see no icon for my 5 static devices.

I notice your link uses https - try using http instead. There's a number of issues with https right now.

Also make sure your JFFS partition is properly mounted. See if this returns any error, over SSH:

touch /jffs/test

 

[RMerlin]

Thanks Merlin, maybe I'll be I who are unlucky!

The only changes I made after the update are:
- I plugged a pen formatted FAT32 for compatibility with Android smartphones (before was NTFS) on port USB 3.0 (as a media server);
- I plugged a pen formatted EXT2 (before there not was) on USB 2.0 port (for install adblock).

Make sure the Media Server is enabled, especially if you did a factory default reset. The media server will no longer default to being enabled after a factory default, as this was a security + performance issue. People plugging their hard disks containing their porn stash were unknowingly sharing with everyone else at home.

 

[stevef9432203]

I can confirm beta2 and release of 380.59 don't show any of the content on the usb drive,.

Further digging found that minidlna is terminating with a sql error on files.db in .minidlna

Quickest fix is "find / -name files.db -exec rm {} \;" in a bash shell. Then reboot.

 

[stevef9432203]

I've noticed the same bug in 380.59 beta2. I'm using the same router: AC87U.

Proble is that minidlna ver changed, breaks sql files.db in .minidlna
Need to delete. Fast fix

"find / -name files.db -exec rm {} \;" in shell on router

 

[cowst]

I notice your link uses https - try using http instead. There's a number of issues with https right now.

Also make sure your JFFS partition is properly mounted. See if this returns any error, over SSH:

touch /jffs/test

Oh, JFFS issues again, didn't expect that.
I already had it specifically with version before .58, and most likely it is happening again because I cannot VPN to home with .59, but I can if I revert to .58

I tried enabling http on 8080, same issue.

BTW: I didn't know I could change icons, cannot see it in .58, although the icon background seems the same you have in the connected devices list.
From page source I cannot find the icons, because they are in JS and the server crashes before getting them.
However I can tell you I have 5 static ones, and I flashed .58 so I can see what was assigned to them from the connected devices list:
2 satellite receivers (laptop icon), 1 laptop (laptop icon), 1 XBOX (is disconnected, cannot see now), 1 TVCam (webcam icon)

 

[do85]

On 380.58 thread I posted a message saying that PPTP connection through my RT-AC68U on 380.57/58 was broken. Then I checked out this 380.59 and the problem is fixed. Thanks!!!