Asuswrt-Merlin 3.0.0.4.374.33 Beta 1 available

[RMerlin]

Thank you for your suggestion. I tried to connect on router LAN port 80 TCP and weird lines immediately appeared in the syslog. This is strange as the OpenVPN server should listen ONLY on the WAN-side port 80, not on the LAN side. Am I wrong?

OpenVPN listens to all interfaces, both WAN and LAN. Might be handy for testing to have it listen on the LAN side. Anyway, you'd still get WAN-side connections to port 80.

I am forced to do this as I am traveling a lot, also my daughter is studying abroad, so very often we are in very paranoid environment. Some organisations like universities, companies and hotels are trying to filter everything except TCP port 80 and also are using HTTP proxies, so connecting to my home OpenVPN server could be impossible. The only solution is to configure the OpenVPN server on TCP port 80. This always guarantees connection as at least port 80 is open on every public network. I had several problems in some places when I used standard OpenVPN ports.

I would recommend putting it on port 443 then. At least it will limit the number of unexpected connection attempts, and port 443 should still be open from pretty much anywhere (this is the HTTPS port).

Merlin, do you have any ideas about problems with existing Entware installation described in my previous posts?

No idea short of your device getting mounted on a different mount point. As I said, I don't have that problem with either my AC66 or N66. I do mount each of them through a label rather than their device names however, makes it more stable.

 

[netware5]

..................
No idea short of your device getting mounted on a different mount point. As I said, I don't have that problem with either my AC66 or N66. I do mount each of them through a label rather than their device names however, makes it more stable.

What you mean saying "your device getting mounted on a different mount point"?

BTW I remember that there were an option in WebUI to mount by label, not by device name, but now I can not find it....

 

[Kobold]

Any performance overall improvement from the compiler optimizations?

I've been using mips32r2 firmwares for a couple of months, and no problems such as accidental reboots or freezes have been detected. I haven't made any mesurements of performance of r2 builds against r1 ones I can't say anything about it. But I strongly believe that mips32r2 binaries work faster.
RMerlin, what version number of the wireless driver is shown on your System Info page? Mine still says '6.30.102.9 (r366174)'. If memory serves this version number was used in early 372 builds or even in the older ones.

 

[matcher]

Thanks for the feedback, so far you're the first user that tells me he is actually using it. Any particular problem with it so far? I was worried that its performance might not be optimal for anyone outside of Russia. Having it limited to only the children's computer might not make it a major issue at least.

If it's working well, I might just promote it to the main build.

Hi Merlin,

First thanks very much for your dedication to providing these firmwares to us.

I am currently on 3.0.0.4.374.32_0-dwrpyd and I also use Yandex.DNS. I have my kids computers/devices on Family Mode and I find it works very well. It is very easy to setup and I have not seen any issues (performance or otherwise) from using it.

I hope you will continue to include it in future releases.

Thanks,

 

[PrivateJoker]

Hi Merlin,

First thanks very much for your dedication to providing these firmwares to us.

I am currently on 3.0.0.4.374.32_0-dwrpyd and I also use Yandex.DNS. I have my kids computers/devices on Family Mode and I find it works very well. It is very easy to setup and I have not seen any issues (performance or otherwise) from using it.

I hope you will continue to include it in future releases.

Thanks,

How do you find the speed of Yandex?

What are the pros/cons of it compared to OpenDNS?

If you're curious about hard benchmark #'s, Steve Gibson has a cool little tool on his site (I think it's Win only) it will compare your DNS resolution times to all the big, and some not so big ones.

https://www.grc.com/dns/benchmark.htm

 

[Deleted member 22229]

The signal level improvement that I saw was on an RT-N66U. Signal level was measured in my "standard" location with a Galaxy Note 2 running WiFi Analyzer. I've always gotten -69 dBM in that location +/- about 2 dBM. With the new beta, I get -64 dBM +/- 2 dBM. I watched it for about 2 minutes like I usually do and I've never gotten a signal level that high in that location. This was running 2.4 Ghz.

Mike

Im running the new bata as well and my roku box went from Good signal stats to Excellent full bars and the roku box don't move so there is an improvement on at least the 5 G side.

 

[RMerlin]

I've been using mips32r2 firmwares for a couple of months, and no problems such as accidental reboots or freezes have been detected. I haven't made any mesurements of performance of r2 builds against r1 ones I can't say anything about it. But I strongly believe that mips32r2 binaries work faster.
RMerlin, what version number of the wireless driver is shown on your System Info page? Mine still says '6.30.102.9 (r366174)'. If memory serves this version number was used in early 372 builds or even in the older ones.

I got the same version number on my RT-AC66U, however you must also look at the build date, as Asus can modify the driver, in which case the version number won't change (that version is Broadcom's original code).

 

[RMerlin]

What you mean saying "your device getting mounted on a different mount point"?

BTW I remember that there were an option in WebUI to mount by label, not by device name, but now I can not find it....

If a device has a label, then it will automatically mount based on the label rather than the device name.

By different mount point I mean a different /mnt/xxx path. If you weren't mounting by label, then it might have been mounting as /mnt/sda1/ , and now it got swapped to /mnt/sdb1 for example. The "mount" command over ssh/telnet will show you what it's using.

 

[RMerlin]

Looks like there might not be any experimental release for the next final build, if I end up moving DualWAN, Repeater and Yandex back to the main release (and now that Asus switched to SDK6 for the RT-N66U, that means no special pixie dusted build either).

That makes me happy. Means five less firmwares for me to compile

Thanks everyone who confirmed that Yandex was working well. I was wondering if that nifty implementation Asus did couldn't be adapted to OpenDNS...

 

[HawkInOz]

Looks like there might not be any experimental release for the next final build, if I end up moving DualWAN, Repeater and Yandex back to the main release (and now that Asus switched to SDK6 for the RT-N66U, that means no special pixie dusted build either).

Sounds like Beta 2 may be based on the forthcoming 3.0.0.4.374.720 GPL code, and include these three components as well as the SDK6 wireless drivers. Fair statement? Thanks!

 

[CivHunter]

Per IP Open DNS filtering

Looks like there might not be any experimental release for the next final build, if I end up moving DualWAN, Repeater and Yandex back to the main release (and now that Asus switched to SDK6 for the RT-N66U, that means no special pixie dusted build either).

That makes me happy. Means five less firmwares for me to compile

Thanks everyone who confirmed that Yandex was working well. I was wondering if that nifty implementation Asus did couldn't be adapted to OpenDNS...

I have been using OpenDNS to filter categories I do not want my kids to have access to and it works well. The down side is I have kids of various ages and this filtering is for my entire LAN. It would be nice to be able to adjust OpenDNS filtering by IP so my teenagers could have fewer restrictions than the little guy. So including that would be awesome!

 

[JGrana]

Minidlna running out of memory

Hi,
Running 374.33 Beta 1 on AC66U. Had been running 374.32.

I have minidlna setup to create it's database on an EXT3 partition, it's media directory on a FAT partition. The media directory contains roughly 12M of music, 280M of photos and 211M of videos.

I have a swap partition mounted, 32M in size. "Free" typically shows at most 2M used.

Everything before Beta 1 has run fine. Minidlna indexes and builds it's database without issues.

With 374.33 Beta 1, minidlna bails out roughly 1/2 way through the Photos indexing (it has already finished the Music). The long error sequence starts with a number of malloc fails:

image_utils.c:503: warn: malloc failed

Then sql reports out of memory a number of times:

sql.c:41: error: SQL ERROR 7 [out of memory]

Then another long series of scanner errors:

scanner.c:488: warn: Unsuccessful getting details for /tmp/mnt/bigback/Photos/2007/Cabin/March/IMG_0763

minidlna continues to run, but will only display Music. No photos or videos.
I have rebooted after the initial install/reboot, and the same sequence occurs....

Anyone else seeing minidlna or memory issues?

 

[RMerlin]

Sounds like Beta 2 may be based on the forthcoming 3.0.0.4.374.720 GPL code, and include these three components as well as the SDK6 wireless drivers. Fair statement? Thanks!

Will depend on when I get the new code, and how it looks like. But yes, the current plan is to switch to the new code once it comes out, and also move the pixie dust back into the main built.

 

[RMerlin]

Hi,
Running 374.33 Beta 1 on AC66U. Had been running 374.32.

I have minidlna setup to create it's database on an EXT3 partition, it's media directory on a FAT partition. The media directory contains roughly 12M of music, 280M of photos and 211M of videos.

I have a swap partition mounted, 32M in size. "Free" typically shows at most 2M used.

Everything before Beta 1 has run fine. Minidlna indexes and builds it's database without issues.

With 374.33 Beta 1, minidlna bails out roughly 1/2 way through the Photos indexing (it has already finished the Music). The long error sequence starts with a number of malloc fails:

image_utils.c:503: warn: malloc failed

Then sql reports out of memory a number of times:

sql.c:41: error: SQL ERROR 7 [out of memory]

Then another long series of scanner errors:

scanner.c:488: warn: Unsuccessful getting details for /tmp/mnt/bigback/Photos/2007/Cabin/March/IMG_0763

minidlna continues to run, but will only display Music. No photos or videos.
I have rebooted after the initial install/reboot, and the same sequence occurs....

Anyone else seeing minidlna or memory issues?

Can you confirm that the swap does get used at all?

Also it's possible that you have a corrupted file on your disk that causes the scanner to crash. 374.33 Beta 1 has a newer version of minidlna in which Asus implemented a function to generate a cache of thumbnails - could also be that those thumbnails are filling up memory.

 

[JGrana]

Thanks for the quick response. I see swap being added in dmesg. From what I recalled, I did a "free" and saw swap at 0, it is now showing 2180. It looks like it is being used.

I also just checked the partition type of the Media dir - it's NTFS, not FAT.

Might be some corruption. Seems to always get to the same general area of the Photos directory (2007). I will try some experiments. One idea is to plug the disk into a Windows machine and run the disk check. Is there an fsck for ntfs mounted disks? I can unmount it and try that from work (Putty ;-)

 

[JGrana]

Ok, removed USB drive from AC66U, plugged into Windows laptop and ran disk check (Fix file system errors checked). Run a second time.

Plugged back into AC66U, malloc failed around the same point of re-indexing...
Free showed swap around 3M.

Then, installed minidlna from Entware. Rebooted router (then realized I needed to copy the minidlna confing to /opt/etc. Rebooted again.

minidlna log showed it running minidlna ver 1.0.25 (versus the 1.1.0 built into 374.33 b1).

Ran fine, indexed everything.

Soooo.... does indeed look like minidlna 1.1.0, when presented with many photos will likely crash, running out of resources.

I will stay with 1.0.25 for now. In fact, not sure how to revert back to 1.1.0)

Merlin - let me know if you want me to do anything.

 

[Brazilian Joe]

Is there a way to 'mess up' with the interface without having to compile the whole firmware?

Like, if I enable jffs and/or add a HDD, is there a directory where it looks for a js file to load and extend the page functionality?

I can elaborate more on the idea if it is interesting, and also lend some coding hand.

 

[RMerlin]

Ok, removed USB drive from AC66U, plugged into Windows laptop and ran disk check (Fix file system errors checked). Run a second time.

Plugged back into AC66U, malloc failed around the same point of re-indexing...
Free showed swap around 3M.

Then, installed minidlna from Entware. Rebooted router (then realized I needed to copy the minidlna confing to /opt/etc. Rebooted again.

minidlna log showed it running minidlna ver 1.0.25 (versus the 1.1.0 built into 374.33 b1).

Ran fine, indexed everything.

Soooo.... does indeed look like minidlna 1.1.0, when presented with many photos will likely crash, running out of resources.

I will stay with 1.0.25 for now. In fact, not sure how to revert back to 1.1.0)

Merlin - let me know if you want me to do anything.

Might be worth seeing if it's a specific photo causing the issues. Try scanning only half of the photos at a time, until you can isolate it to a specific photo (or album).

Otherwise, it could also be a bug introduced in the new thumbnail code that Asus has added, not necessarily 1.1.0 itself.

 

[RMerlin]

Is there a way to 'mess up' with the interface without having to compile the whole firmware?

Like, if I enable jffs and/or add a HDD, is there a directory where it looks for a js file to load and extend the page functionality?

I can elaborate more on the idea if it is interesting, and also lend some coding hand.

Copy the /www files somewhere (USB disk, for example), then mount bind on top of it:

mount -o bind /mnt/USBDISK/www /www
service restart_httpd

That will make httpd start using the files from the www folder on your USB disk. That's how I usually work when doing some webui development.

 

[desides]

Does this release include the newer N66U driver that was removed from prior builds?