ASUS – DSL-AC88U / DSL-AC3100 Firmwares (beta's) etc. Looks like a new Beta Release for the DSL-AC88U is available, this version has Telnet enabled and improved 2.4GHz wifi (I believe). DSL-AC88U_v9.10.06_build581_debug.w.zip https://www.asuswebstorage.com/navigate/s/604D61534F2541D1B3CE1CC031A697E4Y UK Sentinel In a completely...

[UK Sentinel]

UK Sentinel wrote:

jetspeed wrote:

Curious to see what happens to your RT-AC86U after a few days. It looks like the RAM usage increases 1% per day on the DSL-AC88U. I have web history off. Does the unit store more of the system log than we are able to view? I know the unit caches DNS lookups, but is it worth like 6MB/day? Perhaps someone who doesn’t use the internet everyday can test, but don’t like our chances of find someone to do that

I originally rebooted my RT-AC86U as I noticed RAM used was 435 MB (85%) Routers typically use RAM for Cache/Buffers, so it could well store data such as DNS lookups as cache this would make router more efficient/quicker with that service etc. Will report back after weekend with RAM usage for RT-AC86U, as I also am keen to see if it ever runs out of RAM?

RAM usage for RT-AC86U is now 386MB (75%) and was previously 373MB (73%) so RAM usage is increasing by say 1% per day (ish)

In a completely sane world, madness is the only freedom (J.G.Ballard).

You need to login in order to vote

[pavlf]

Well, I did connect a few more devices by wifi over the weekend, and did find that usage went up. It’s now at 225mb, was at 220, two days ago. It’s very slow, but it’s definitely happening. It’s definitely a waiting game. Quite what will happen if after 30 days it reboots, I’m not sure. Seems like maybe ASUS have moved on. Live with it or sell it I suppose.

You need to login in order to vote

[KitsuneDave]

11 days since my DLM got reset and back up to my 330Mbps Down, 50Mbps Up, connection and has been stable on the official 585 firmware. Ive learned to live with the 88u. Ive had no issues with my speed, or with high cpu usage/mem usage when under high load etc. Since the majority of issues revolve around the ai protection ive just left it disabled as its more troublesome than its worth when enabled. To be honest if any user has the knowledge and common sense its not really needed. Another reason why my speed plummeted i think is because with ai on, the Nat acceleration gets disabled for precise packet inspection. Users that are below 100Mbps might not be as affected as me when its on.

Memory usage usually hovers around 37%, max ive seen is 42% and thats with 3 wired desktops, 16 wireless devices under heavy usage. Cpu usage is very minimal too. No dropouts on the wireless either.

Things ive changed:

2.4 & 5ghz – airtime fairness disabled, preamble type short, All beamforming disabled

Upnp disabled

Dos protection enabled

Static IP’s for everything

 

 

 

You need to login in order to vote

[oxygen8]

jetspeed wrote:

Curious to see what happens to your RT-AC86U after a few days. It looks like the RAM usage increases 1% per day on the DSL-AC88U. I have web history off. Does the unit store more of the system log than we are able to view? I know the unit caches DNS lookups, but is it worth like 6MB/day? Perhaps someone who doesn’t use the internet everyday can test, but don’t like our chances of find someone to do that

My RT-AC86U has been up for 7 days and 12 hours and the RAM oscillates between 66% and 68% usage with AiProtection, Firewall, Parental control, etc, but; I’m using RMerlin 384.7_2 firmware.

I’m even using Smart connect to push clients to the 5GHz and it is working like a charm.

You need to login in order to vote

[oxygen8]

And I just updated to RMerlin 384.8 which was released last night ;-)

You need to login in order to vote

[UK Sentinel]

oxygen8 wrote:

And I just updated to RMerlin 384.8 which was released last night

thanks for the Heads-Up on that, I need to update mine also

For those that are interested, changelog below for reference:

384.8 (2-Dec-2018)
– NOTE: Asus has put the RT-AC56U on their End of Life
list, meaning no further firmware releases from
them.  Since it’s impossible for me to support
models without matching GPL releases from Asus,
I also have to retire the RT-AC56U.  384.6 is
the final release for that model.

– NOTE: The RT-AC3200 and RT-AC87U are not supported by this
release, Asus hasn’t released any updated code yet for
these models.

– NEW: Added RT-AX88U support (based on GPL 384_4736).
– NEW: Merged with GPL + binary blobs from 384_32799 (all
supported models except RT-AX88U)
– NEW: Add LZ4 V2 option to OpenVPN compression
(more effective at handling already compressed
data)
– NEW: Added “extend” support to SNMP.
– NEW: Added CleanBrowsing to DNSFilter supported services.
– NEW: Webui HTTP LAN port can now be changed from the default 80.
– NEW: Added support for the Netfilter TEE target.
– CHANGED: Removed watchdog from OpenVPN clients, to avoid
conflicting with more advanced configurations.
– CHANGED: Vsftpd TLS mode will now reuse the web server
certificate (including any Let’s Encrypt generated
one).
– CHANGED: SSL crypto/cipher hardening for httpd (themiron)
– CHANGED: Syslog will now ignore bwdpi debug output (themiron)
– CHANGED: Reworked Wireless Log page, adding a new button to
view low-level details (what stock firmware shows
on its Wireless Log page), and removed redundant
option to display DFS channel details.
– CHANGED: Update dnsmasq to 2.80-11-g59e4703 (themiron)
– CHANGED: Updated nettle to 3.4
– CHANGED: Updated net-snmp to 5.8
– CHANGED: Updated openssl to 1.0.2q
– CHANGED: Migrated /jffs/ssl/* content to /jffs/.cert (to
share the same folder used by Asus stock)
– CHANGED: Re-enabled WTFast on non-HND models (curl-related
crash has been fixed).  This is still untested.
– CHANGED: Updated CA bundle to October 17th 2018 version.
– CHANGED: Support search domains pushed by a remote OpenVPN
server
– FIXED: UOPNP port forwarding not working in CGNAT/double NAT
scenario even if proper ports were forwarded upstream.
– FIXED: Pages based on table.js (like the port trigger one)
would fail to work properly under Firefox
(Michael Ziminsky)
– FIXED: Dnsmasq issues when running in non-router mode
(John Bacho)
– FIXED: Routing issues when in non-router mode (John Bacho)
– FIXED: Bug in curl that could cause some applications to
crash on non-HND models
– FIXED: IFTTT failing to start on non-HND models (caused by
curl issue).
– FIXED: Webui could complain about port 8080 being reserved for
http WAN port (which is no longer supported)
– FIXED: Cannot change image for device with a vendor name
containing an apostrophe (like Micro-Star int’l)
(Asus bug)
– FIXED: OpenVPN client download was capped by Adaptive QOS
upload limit (fix devised by FreshJR)
– FIXED: OpenVPN custom config might be lost after a reboot
on the RT-AC86U.

In a completely sane world, madness is the only freedom (J.G.Ballard).

You need to login in order to vote

[jetspeed]

Day 24: 339MB used (74%), Build 585, AiP OFF

UK Sentinel wrote:

RAM usage for RT-AC86U is now 386MB (75%) and was previously 373MB (73%) so RAM usage is increasing by say 1% per day (ish)

This is the same behaviour I’m experiencing with my DSL-AC88U. Looks like they share some common code base. oxygen8 doesn’t have problems with Merlin FW so it’s also seemingly an ASUS thing?

KitsuneDave wrote:

Static IP’s for everything

I’ll might give this a try when/if my unit reboots. I currently have about 38 devices on DHCP. Maybe it’s writing to some log file.

You need to login in order to vote

[KingfisherUK]

Interesting find re: the IPv6 issue I (and others) have been having on BT VDSL; I now think the issue may lie with the chipset itself.

I’ve got my hands on a Billion 8900AX-1600 R2 to play with which is (I believe) based on the same Broadcom chipset as the DSL-AC88U and weirdly enough it seems to have exactly the same issue – either via the build in DSL modem OR external BT Openreach modem, it detects the PD as 64, not 56 as it should be.

Could just be coincidence…

You need to login in order to vote

[UK Sentinel]

I understand that from Broadcom (DSL and Wifi) and TREND etc, the code (source code) or as RMerlin calls the code (Blobs) arrives to ASUS from various vendors,  then ASUS  join the codes (blobs)  together, so this is where the issues could be with firmware’s, Billion/Netgear etc. all have the blobs of code from chipset vendors.

Billion are usually very good when it comes to DSL related kit.

 

In a completely sane world, madness is the only freedom (J.G.Ballard).

You need to login in order to vote

[KingfisherUK]

UK Sentinel wrote:

Billion are usually very good when it comes to DSL related kit.

Which is exactly why I was surprised to find the same issue.

You need to login in order to vote

[UK Sentinel]

KingfisherUK wrote:

UK Sentinel wrote:

Billion are usually very good when it comes to DSL related kit.

Which is exactly why I was surprised to find the same issue.

Maybe the same blobs are shared amongst differing DSL Router vendors?

In a completely sane world, madness is the only freedom (J.G.Ballard).

You need to login in order to vote

[KingfisherUK]

Same conclusion I came to, which suggests it’s an inherent “feature” of the chipset itself. If so, that basically renders any device based on it useless on BT VDSL services.

You need to login in order to vote

[UK Sentinel]

Just had a look online, and seems the Netgear D7000-100UKs also uses the BCm63138 chipset and only supports /64

I forget, do you have a Home HUB etc? and does this work with /56 ?

In a completely sane world, madness is the only freedom (J.G.Ballard).

You need to login in order to vote

[KingfisherUK]

Got a BT SmartHub (v1) and a HG612 Openreach modem which I can use with my old RT-AC66U and both pickup /56 and work perfectly.

Not good news about the Netgear, really does look like it’s a no-go with this chipset. Oh well, guess the one-box solution is out the window then…

You need to login in order to vote

[UK Sentinel]

I will have another chat with ASUS – the BT SmartHub v1 uses a Broadcom 63137 DSL chipset, which is a less advanced version compared to BCM63138 – but still is (1 GHz, 2 cores)

Update: I spoke with ASUS and they are looking into the IPV6 /64  /56 challenges (my wording not there’s)  ;-)

In a completely sane world, madness is the only freedom (J.G.Ballard).

You need to login in order to vote

[Author]

Posts