Release Asuswrt-Merlin 386.7 is now available for all models

[geobernd]

The LAN link speed on my AC66U_B1 (Merlin Firmware AC68U 386.7) drops to 100Mbps from 1Gbps and stays at 100Mbps. If I unplug and then reconnect the LAN cables again, link goes up to 1Gbps for an hour or so and then drops again. Auto Negotiation is OFF on my desktop ethernet settings and set to 1Gbps Full Duplex.

What could be the issue?

As a first I woudl try another Ethernet cable: I had the same on my RT-AC68P on the WAN to Cable Modem (SB6183) link a while back and was sure it was a port going bad since it happened randomly after a while. On a hunch I replaced the cable and the problem went away and never returned...

 

[Wrkdbf_Guy]

Dirty upgrade from 386.5 to 386.7 on my main RT-AC68U router. Everything's going along swimmingly.

BTW yes, I know that router is old and slow relative to today's newer models. But I only have 160 Mbps internet access and it works really well in that scenario. Dependable as all get out too (knocking loudly on wood).

 

[d3viLD00m]

As a first I woudl try another Ethernet cable: I had the same on my RT-AC68P on the WAN to Cable Modem (SB6183) link a while back and was sure it was a port going bad since it happened randomly after a while. On a hunch I replaced the cable and the problem went away and never returned...

I will definitely try this, but the probability of all the 3 cables going faulty at the same time is a bit unlikely. I bought a 5-pack CAT7 cables a year back, to use on both WAN/LAN. WAN stays at 1 Gbps, but not the LANs.

 

[L&LD]

Yes, try a different Cat5e Ethernet cable (and from a different manufacturer too) and be sure you put the PC back to Auto negotiation too.

And, don't forget to reboot everything if you want things stable.

 

[capncybo]

@Sneakytiki I also should have asked, "Aggregate to WHAT?" meaning are you combining 2 router's Ethernet ports & connecting them with:
-An IP-Provider's Cable Modem
-The AiMesh Node
I ask this because... I often read Keep the AiMesh Nodes on stock firmware but...
With all the significant changes to Asus/Merlin firmware as of late.
For an advanced feature such as Network Aggregation...
Perhaps the Nodes need to be on the same firmware version as the router?
Just brainstorming.

 

[Sneakytiki]

Sorry for all your troubles, & I can sympathize with today's families being so reliant on technology... any network disruptions can be a real challenge. What I had completely missed via our earlier conversation & I'm not sure if RMerlin caught it himself was... YOU ARE USING AGGREGATION. I think it's a safe assumption that MOST people are NOT using this feature. Hence I can see how an aggregation-type problem could have been missed. Good job identifying & reporting it.

Yeah, no worries, its always a blast trying to figure out the primary cause of something. Aggregation is the issue with most of the problems, at least I think it is. There is something wonky with wireless and ethernet at times too, especially with 160mhz and maybe full ethernet speeds, but its not a deal breaker since we would never saturate that. We have never used our entire connection (not sure the router can with TrendMicro and adaptive QoS) except for very rare days when we are are streaming, conferencing and gaming/downloading.

Initially, I thought it was something with diversion or something sticking in JFFS but even disabling that and formatting the other with a complete fresh start with no USB, we still had issues. Plus, reboots and errors only using one cable between modem and router, mostly wireless issues.

We also use TrendMicro and Adaptive/Flex QoS for the kids because they click on things and since Comcast upload speeds are trash and zoom meetings/school/work. We like diversion, stop devices calling home, and adblocking and I mess with Skynet and Unbound for fun and out of curiosity. Plus other smaller scripts like ntpMerlin, menus, etc.

 

[John Fitzgerald]

We also use TrendMicro and Adaptive/Flex QoS for the kids because they click on things and since Comcast upload speeds are trash and zoom meetings/school/work. We like diversion, stop devices calling home, and adblocking and I mess with Skynet and Unbound for fun and out of curiosity. Plus other smaller scripts like ntpMerlin, menus, etc.

Trendmicro, QoS and Diversion ***(with pixelserv-tls) are known to clash.

 

[Sneakytiki]

@Sneakytiki I also should have asked, "Aggregate to WHAT?" meaning are you combining 2 router's Ethernet ports & connecting them with:
-An IP-Provider's Cable Modem
-The AiMesh Node
I ask this because... I often read Keep the AiMesh Nodes on stock firmware but...
With all the significant changes to Asus/Merlin firmware as of late.
For an advanced feature such as Network Aggregation...
Perhaps the Nodes need to be on the same firmware version as the router?
Just brainstorming.

Firmware on nodes (AC3100) did not make a difference and aggregation is between modem (CM1150V) and router (AX88U) only. It either sits there attempting to get an address and then reboots or stalls with an error (DHCP) or it says network cable disconnected.

Ethernet cable or wireless mesh did not make a difference either. Changing query frequency didn't do anything. Power cycling router or modem did not make a difference. Resetting them did not make a difference. Disabling AIMesh did not make a difference. I could only reach the cable modem with an ethernet cable, going through the router was not possible with 192.168.100.1.

I initially thought the modem was having issues, lots of uncorrectables, something I had not seen since switching from a Puma based modem a few years ago. It seems fine now.

 

[Sneakytiki]

Trendmicro, QoS and Diversion ***(with pixelserv-tls) are known to clash.

Yes, I did see some issues with that when I was searching through the site, it was definitely on my list with the firmware upgrade. We rebuilt the entire network, new name, new passwords, default settings, just merlin on the router and it was still having issues.

I do want to try AdGuard on the router, just need to set some time aside to read through it all. Wanted to try a Pihole too but I can't find any Pi's in my area, everyone is still sold out.

ETA: Removed Diversion, added AdGuard, working great, even w/ Unbound. Same issues as before though, can JFFS or something else stick after a format and factory reset? I really have no idea what it is unless its something buried in the wireless and other updates.

 

[bluepoint]

Rt-AX88U owners do you see these continous logs after three days? I have no Aimesh setup just one addon(diversion lite) and guest clients.

Jun 26 10:30:54 kernel: CONSOLE: 171145.317 wlc_ap_authresp: status 0
Jun 26 10:30:54 kernel: CONSOLE: 171145.322 wlc_ap_process_assocreq_done status 0
Jun 26 10:30:54 kernel: CONSOLE: 171145.325 iov:SCB_DEAUTH
Jun 26 10:30:54 kernel: CONSOLE: 171145.327 tx:prep:802.1x
Jun 26 10:30:54 kernel: CONSOLE: 171145.334 wl0.2: wlc_send_bar: for d4:a6:51:xx:xx:xx seq 0x1 tid 5
Jun 26 10:30:54 kernel: CONSOLE: 171145.337 tx:prep:802.1x
Jun 26 10:30:54 kernel: CONSOLE: 171145.339 iov:SCB_AUTH
Jun 26 10:30:54 kernel: CONSOLE: 171145.365 wl0.2: wlc_send_bar: for d4:a6:51:xx:xx:xx seq 0x1 tid 6
Jun 26 10:30:55 kernel: CONSOLE: 171146.095 wl0.2: wlc_send_bar: for 84:e3:42:xx:xx:xx seq 0x16 tid 0
Jun 26 10:30:57 kernel: CONSOLE: 171147.547 wl0.2: wlc_send_bar: for d4:a6:51:xx:xx:xx seq 0x1 tid 0
Jun 26 10:30:57 kernel: CONSOLE: 171147.667 wl0.0: wlc_send_bar: for f0:72:ea:xx:xx:xx seq 0xc3 tid 3
Jun 26 10:30:57 kernel: CONSOLE: 171148.085 wl0: wlc_ampdu_recv_addba_resp: 54:60:09:xx:xx:xx: Failed. status 37 wsize 64 policy 1
Jun 26 10:30:59 kernel: CONSOLE: 171150.288 wl0.2: wlc_send_bar: for 84:e3:42:xx:xx:xx seq 0x18 tid 0
Jun 26 10:31:00 kernel: CONSOLE: 171151.071 wl0: dc:91:bf:xx:xx:xx: addba timed out 0
Jun 26 10:31:00 kernel: CONSOLE: 171151.358 wl0.2: wlc_send_bar: for dc:91:bf:xx:xx:xx seq 0x173 tid 0
Jun 26 10:31:01 kernel: CONSOLE: 171151.485 wl0.0: wlc_send_bar: for 54:60:09:xx:xx:xx seq 0x1 tid 3
Jun 26 10:31:14 kernel: CONSOLE: 171164.856 wl0.2: wlc_send_bar: for dc:91:bf:xx:xx:xx seq 0x17c tid 0
Jun 26 10:31:18 kernel: CONSOLE: 171169.113 wl0.2: wlc_send_bar: for dc:91:bf:xx:xx:xx seq 0xa tid 3
Jun 26 10:31:19 kernel: CONSOLE: 171170.196 wl0.2: wlc_send_bar: for 84:e3:42:xx:xx:xx seq 0x19 tid 0
Jun 26 10:31:21 kernel: CONSOLE: 171172.212 wl0.2: wlc_send_bar: for dc:91:bf:xx:xx:xx seq 0x18c tid 0
Jun 26 10:31:25 kernel: CONSOLE: 171175.491 wl0.0: wlc_ampdu_resp_timeout: 54:e0:19:xx:xx:xx: tid 0 cleaning up resp tid waiting for seq 0x50 for 200 ms
Jun 26 10:31:28 kernel: CONSOLE: 171178.374 wl0.2: wlc_send_bar: for dc:91:bf:xx:xx:xx seq 0x19c tid 0
Jun 26 10:31:30 kernel: CONSOLE: 171180.985 wl0: wlc_ampdu_recv_addba_resp: d4:a6:51:xx:xx:xx: Failed. status 37 wsize 16 policy 1
Jun 26 10:31:30 kernel: CONSOLE: 171180.985 wl0: wlc_ampdu_recv_addba_resp: d4:a6:51:xx:xx:xx: Failed. status 37 wsize 16 policy 1
Jun 26 10:31:30 kernel: CONSOLE: 171180.986 wl0: wlc_ampdu_recv_addba_resp: d4:a6:51:xx:xx:xx: Failed. status 37 wsize 16 policy 1

@RMerlin
These are the logs I was referring to from the already closed thread. Waited for awhile to post it until it comes out. New debug log from this GPL?

 

[Aiadi]

@aidil, did you try clearing your browser cache? Did you try accessing the router via the browser's safe mode (i.e. with all extensions disabled)?

After doing the full reset, are you importing an old backup config file?

Try using a different browser too (which one are you using?).

Thank you for your reply but no joy unfortunately. Cleared all settings on chrome, edge and brave browsers to no avail. I am a bit at a loss.

 

[Tvlz]

Hi, I noticed something wrong with IPv6 page on my RT-AX86U after update to 386.7; indeed, when I was on 386.5_2 my IPv6 page looks like this:

Anyhow, my ISP provide me a 48 bit length IPv6 range of addresses; what I'm doing wrong right now? Why I stepped down from 48 to 64 bit?

But why before, with 386.5, I had a /48 prefix for Lan too?

Nothing, IPv6 Prefix Delegation was broken before now it is fixed.
Your ISP gives you a /48 WAN prefix which is 64K /64 subnets(18 quintillion IP addresses per /64) for use on your LAN, the GUI displays only 1 of those /64 subnets.

@Tvlz must have had plans to manually assign other IPv6 subnets from the larger prefix, because even dnsmasq is told to only advertise a hard-coded 64 via stateless RAs:

A /64 prefix is the standard size used with IPv6 Prefix Delegation, anything larger(/60 /56 /48) and you are talking about subnets of /64.

How do you configure the network to make use of the extra subnets available in a /48? Seems challenging.

Right now it is a manual process it needs to be scripted but not too difficult to setup, due to asus not having a GUI for it.

It is, but still, why drop this feature if it'd be available?

Nothing was dropped just fixed.

 

[John Fitzgerald]

Thank you for your reply but no joy unfortunately. Cleared all settings on chrome, edge and brave browsers to no avail. I am a bit at a loss.

EDIT: see post #300

Try doing a GUI "Factory Default Reset". From the Fresh Instant, re-flash the firmware, then directly doing another GUI Reset, before configuring your settings. I would do the settings from scratch again. (yes it's a PITA)

 

[matthew_eli]

Nothing, IPv6 Prefix Delegation was broken before now it is fixed.
Your ISP gives you a /48 WAN prefix which is 64K /64 subnets(18 quintillion IP addresses per /64) for use on your LAN, the GUI displays only 1 of those /64 subnets.


A /64 prefix is the standard size used with IPv6 Prefix Delegation, anything larger(/60 /56 /48) and you are talking about subnets of /64.

Right now it is a manual process it needs to be scripted but not too difficult to setup, due to asus not having a GUI for it.

Nothing was dropped just fixed.

Thanks and sorry: I now realized I know nothing on IPv6! Eh eh

 

[AlbumMan]

RT-AX3000 dirty flash from 386.5_2 --> 386.7 was smooth but noticed a couple of lines that never appeared in any of the previous logs that are set to Notice / Warning:

May 5 00:05:07 kernel: gluebi (pid 1): gluebi_resized: got update notification for unknown UBI device 0 volume 0
(Did a reboot after 30 minutes and that line was not there)

May 5 00:05:07 kernel: hub 1-0:1.0: config failed, hub doesn't have any ports! (err -19)
(This line was there again after the reboot)

After 12 hours now everything seems okay so I'm guessing nothing is wrong but did notice RAM usage which has always been 47% on all previous firmware is now 54%.

 

[Kitsap]

RT-AC68U dirty upgrade from 386.5 via Ethernet and Firefox successful. Device configured as an access point and IPv6 works fine. Running smooth for several hours so far. Thank you!

 

[Kendo]

Clean upgrade from 386.5_2 on AX86U, for a weird reason at first the 2.5Gb connection was selected instead of the 1Gb WAN port and couldn't change it.
After the second reset it worked.
First issue of this kind here.
Thank you for the new fw.

 

[L&LD]

A second reset? Do you mean reboot?

 

[Kendo]

No, actually did a reset again and then successfully detected the cable being in the wan port.

 

[RMerlin]

I have to disable the whole ai protection malicious website blocking is enabled by default regardless of me turning it off or on again which is hella annoying with false positives, it feels like the router is trying to replace job of my isp forcing me to use vpn

What router model? How are you determining it's AiProtection's Malicious Website blocking and not something else (like your DNS server)?

Thank you. It is really odd but that option in administration --> system appears to show up very momentarily when I click the system tab on my GT-AX6000 and then quickly disappears into thin air!!! I am running latest AsusMerlin after doing a full reset. Is this some sort of a bug?

The GT-AX6000 does not use JFFS2, they use UBIFS. The format option is not available on this platform.

The LAN link speed on my AC66U_B1 (Merlin Firmware AC68U 386.7) drops to 100Mbps from 1Gbps and stays at 100Mbps. If I unplug and then reconnect the LAN cables again, link goes up to 1Gbps for an hour or so and then drops again. Auto Negotiation is OFF on my desktop ethernet settings and set to 1Gbps Full Duplex.

What could be the issue?

Enable auto-negotiation, it's mandatory for Ethernet Gigabit, and try replacing the cable. Rate dropping usually indicate an unstable connection.