RT-AX86U with latest version 3004.388.8_2 always crash and reboot

[wuwentao]

I'm using AX86U and upgrade to latest 3004.388.8_2 firmware.
device always crash and all the wireless client disconnect.

I have add a usb fan in the back of this AX86U, so the hardware temperature should not any problem:

RT-AX86U	Current Version : 3004.388.8_2

BCM490x - B53 ARMv8 revision 0 (Cores: 4) @ 1800 MHz

Temperatures	2.4 GHz: 40°C - 5 GHz: 42°C - CPU: 61°C

about 50 clients, and most of the client is 2.4G IoT devices, about 10 wired clients for android TV, apple TV and PC devices, but most of my devices only connected to the wireless, and no more traffics.
only 1 iPAD and 1 Mobie phone have online video and have more traffic.

for router feature:
1. pppoe
2. static dhcp ip address with dnsmasq config
3. ddns custom script
4. Traffic Analyze ( I have disabled it today)
5. two virtial server/port forwarding( as I'm using ddns)
6. manual start a tailscale service with custom script in jffs, and only used when I'm out of home for homeassistant web connection and ssh service.
7. enabled

that's all the service that I have enabled in my AX86U.

but the wireless service is really not stable and I can found more crash log from the syslog.
when I just save it from the web buttom [Save], seems it can't save the crash log, so I just copy all the output from the system log window and paste it to another log file again.

for example, I can found a crash in yesterday from the log(copy_pasted log) as below:

start from Aug 25 21:57:02 and time become May 5 13:05:12, after crash and reboot, ntp should works, then time become to normal after Aug 25 21:58:50 ntpd: Initial clock set

Aug 25 21:56:42 hostapd: eth6: STA 36:47:d4:06:e4:14 IEEE 802.11: disassociated
Aug 25 21:56:58 wlceventd: wlceventd_proc_event(685): eth6: Auth 36:47:D4:06:E4:14, status: Successful (0), rssi:0
Aug 25 21:56:58 hostapd: eth6: STA 36:47:d4:06:e4:14 IEEE 802.11: associated
Aug 25 21:56:58 wlceventd: wlceventd_proc_event(722): eth6: Assoc 36:47:D4:06:E4:14, status: Successful (0), rssi:-36
Aug 25 21:57:01 hostapd: eth6: STA 36:47:d4:06:e4:14 RADIUS: starting accounting session 1232F86808FE0A52
Aug 25 21:57:01 hostapd: eth6: STA 36:47:d4:06:e4:14 WPA: pairwise key handshake completed (RSN)
Aug 25 21:57:01 dnsmasq-dhcp[2684]: DHCPDISCOVER(br0) 36:47:d4:06:e4:14
Aug 25 21:57:01 dnsmasq-dhcp[2684]: DHCPOFFER(br0) 192.168.2.112 36:47:d4:06:e4:14
Aug 25 21:57:02 dnsmasq-dhcp[2684]: DHCPREQUEST(br0) 192.168.2.112 36:47:d4:06:e4:14
Aug 25 21:57:02 dnsmasq-dhcp[2684]: DHCPACK(br0) 192.168.2.112 36:47:d4:06:e4:14 360-doorbell-R5MAX
May  5 13:05:12 syslogd started: BusyBox v1.25.1
May  5 13:05:12 kernel: klogd started: BusyBox v1.25.1 (2024-07-31 19:48:48 EDT)
May  5 13:05:12 crashlog: LOG
May  5 13:05:12 kernel: Booting Linux on physical CPU 0x0
May  5 13:05:12 crashlog: <4>^[[0;36;44mPathstats allocated 2048 bytes^[[0m
May  5 13:05:12 kernel: Linux version 4.1.52 (merlin@ubuntu-dev) (gcc version 5.5.0 (Buildroot 2017.11.1) ) #2 SMP PREEMPT Wed Jul 31 20:51:42 EDT 2024
May  5 13:05:12 kernel: CPU: AArch64 Processor [420f1000] revision 0
May  5 13:05:12 crashlog: <4>^[[0;36;44mBroadcom Packet Flow Cache HW acceleration enabled.^[[0m
May  5 13:05:12 kernel: Detected VIPT I-cache on CPU0
May  5 13:05:12 crashlog: <4>Enabled Runner binding to Flow Cache
May  5 13:05:12 kernel: alternatives: enabling workaround for ARM erratum 845719
May  5 13:05:12 crashlog: <4>Initialized Runner Protocol Layer (800)
May  5 13:05:12 crashlog: <4>Broadcom Runner Blog Driver Char Driver v0.1 Registered <309>
May  5 13:05:12 kernel: PERCPU: Embedded 15 pages/cpu @ffffffc03ffb3000 s21400 r8192 d31848 u61440
May  5 13:05:12 crashlog: <3>Lservices Loaded - NComm, Inc.
May  5 13:05:12 crashlog: <3>exLinuxETH Loaded - NComm, Inc.
May  5 13:05:12 kernel: Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 244480
May  5 13:05:12 crashlog: <6>NComm TMS V6.91 Kernel Module loaded.
May  5 13:05:12 kernel: Kernel command line: root=ubi:rootfs_ubifs ubi.mtd=0 rootfstype=ubifs coherent_pool=4M cpuidle_sysfs_switch pci=pcie_bus_safe rootwait
May  5 13:05:12 kernel: PID hash table entries: 4096 (order: 3, 32768 bytes)
May  5 13:05:12 crashlog: <4>PCIe HCD (impl1)^M

and I also found some important crash log as below:

May  5 13:05:12 crashlog: <4>^[[0;33;41m[ERROR pktrunner] pktrunner_wlan_mcast_delete,410: No DHD Stations to be deleted from FWD Table index 0^[[0m
May  5 13:05:12 crashlog: <4>^[[0;33;41m[ASSERT ../../bcmdrivers/broadcom/char/pktrunner/bcm94908/pktrunner_mcast.c] runnerMcast_deactivate,775: mcastFlow_p->result.wlan_mcast_fwd_table_index == RDPA_WLAN_MCAST_FWD_TABLE_INDEX_INVALID^[[0m
May  5 13:05:12 crashlog: <4>BUG: failure at ../../bcmdrivers/broadcom/char/pktrunner/bcm94908/pktrunner_mcast.c:775/runnerMcast_deactivate()!
May  5 13:05:12 crashlog: <0>Kernel panic - not syncing: BUG!
May  5 13:05:12 crashlog: <4>CPU: 2 PID: 1891 Comm: mcpd Tainted: P           O    4.1.52 #2
May  5 13:05:12 crashlog: <4>Hardware name: Broadcom-v8A (DT)
May  5 13:05:12 crashlog: <0>Call trace:
May  5 13:05:12 crashlog: <4>[<ffffffc000087398>] dump_backtrace+0x0/0x150
May  5 13:05:12 crashlog:

the detail can be found with attach log files:
1. ax86u_syslog.txt , use save button to export from web UI , seems missed the crash log
2. ax86u_copy_paste.txt, copy the log from the log display text area and paste it to a text file

 

[Ripshod]

Which firmware did you upgrade from?
Did you factory reset and manually setup (without restoring any configs)?

 

[wuwentao]

Which firmware did you upgrade from?
Did you factory reset and manually setup (without restoring any configs)?

already reset to factory ater upgrade to this firmware (before I post it, some weeks ago), also format the jffs and reinit/reset all the configs/settings, as I'm also thinking maybe some old setting broken it, so I had did it some weeks ago, seems got the same result.
I'm not sure with nvram args. seems there is no any reset option from the web.

and after factory reset, I didn't restore any settings from the old backup, manual add all the settings one by one via web UI, as there is no any specialy configs for me except the dhcp static list.
I just use nvram show |grep dhcp_staticlist to export it. then recovery it with nvram set dhcp_staticlist="xxxx", this is the only one not setting via the web ui.

 

[wuwentao]

just seach the func keyword and error keyword from github repo, not any any useful info, as I'm not familar with c language.
Is there any clue for me to disable some feature or setting to prevent this crash before we got some solution or bug fix?
it should works well with old firmware, device can running with more than 30-60 days without crash or reboot.
I have reboot it and will continue monitor it. if issue still exist, will consider to downgrade to old firmware. as device reboot will cause my tailscale service and network disconnect.

 

[Squall Leonhart]

the usb fan itself can crash it.

 

[wuwentao]

the usb fan itself can crash it.

I'm using 5V/2A power adapter for the usb fan, not plug it to the router, so there is no any usb devices in my router.

 

[Imaginous]

I have 3 of these routers and experience lockups on one of them. This one becomes unreachable after a few weeks.
This one is under light load.

The other two are under medium load and sometimes start disconnecting / reconnecting the clients every few minutes.
After a reboot it's working again for a few days to weeks.

A factory reset didn't help.

After finding this post I downgraded from the 2 to the 0 version again to see if it helps.

Just my 2 cents

 

[wuwentao]

thanks for your info, I will try to downgrade, as it's really not stable, only can works for some days and crash or wifi can't connect again, after reboot, it recovery to normal status.
but I forgot my old firmware version, it maybe 388.6 or 388.5, and it works more stable,I'm not sure the version now.....
as i found there is too many SW upgrade and also include some CVE, so I just upgrde it, once got so many crash, did a factory reset to confirm again.

once new version is stable and i will try to upgrade , thanks

 

[DiGz_Au]

386.5-2 is solid. I've had my ax86u drop all clients a few times now and also have done clean installs and still happens. Will wait and hope for new asus update or go back to said firmware.

 

[Tech9]

386.5-2 is solid.

It was the best Asuswrt-Merlin version for RT-AX86U in 386 series firmware, but >2 years old now. If you still want to use it make sure the router is locked down with no services open to WAN or it's behind your ISP Gateway firewall, if you have one. Otherwise run current stock Asuswrt, it's stable.

 

[Imaginous]

For me 3004.388.8_0 also seems stable since the downgrade.

For me the instability introduced with the 3004.388. 8_2 version.

 

[Ripshod]

also format the jffs

You still have that option available?

 

[learning_curve]

For me 3004.388.8_0 also seems stable since the downgrade.

For me the instability introduced with the 3004.388. 8_2 version.

FWIW I haven't had ANY issues at all with my AT-AX86U, whilst running 3004.388.8_2 i.e. Zero / Nada / Nothing and it was a dirty upgrade too.
Yes, we have different setups, but again FWIW my temps are the same as the OP's, but plus 5℃ on each
No, I don't use (and believe I don't need to use) any additional cooling fan / fans.
Surely this can only be a router config / ISP provided package, related issue?
This firmware release is already proven to me and many other users too.

 

[Imaginous]

I use the routers as access points on 2 locations. On both locations I experience issues.

I don't know the trigger, but after a while it goes haywired.

I only replied to the topic starter I also have strange issues with this version.

 

[meruserasus]

For me 3004.388.8_0 also seems stable since the downgrade.

For me the instability introduced with the 3004.388. 8_2 version.

Same on RT-AX86U Pro, started disconnecting issues for all devices, but merlin said no changes were made to that part of the OS, so... there's nothing to fix

 

[Mogsy]

Noticed a crash yesterday. Checking my log, I had 2 crashes logged. But the normal closed source log. Updated TrendMicro’s signature yesterday.Otherwise this firmware has been solid for me, only the occasional 160MHz drop because of radar signals, I think.

Sep  2 13:22:52 kernel: potentially unexpected fatal signal 11.
Sep  2 13:22:52 kernel: CPU: 0 PID: 1254 Comm: asd Tainted: P           O    4.1.52 #2
Sep  2 13:22:52 kernel: Hardware name: Broadcom-v8A (DT)
Sep  2 13:22:52 kernel: task: ffffffc02407f540 ti: ffffffc0280d0000 task.ti: ffffffc0280d0000
Sep  2 13:22:52 kernel: PC is at 0xf69c1b74
Sep  2 13:22:52 kernel: LR is at 0xf6500018
Sep  2 13:22:52 kernel: pc : [<00000000f69c1b74>] lr : [<00000000f6500018>] pstate: 20010010
Sep  2 13:22:52 kernel: sp : 00000000f6893cd8
Sep  2 13:22:52 kernel: x12: 00000000f650001c
Sep  2 13:22:52 kernel: x11: 0000000000000010 x10: 00000000f6500010
Sep  2 13:22:52 kernel: x9 : 00000000f6893e44 x8 : 0000000000000002
Sep  2 13:22:52 kernel: x7 : 00000000f6500010 x6 : 00000000f6501d38
Sep  2 13:22:52 kernel: x5 : 000000000001f2c8 x4 : 00000000f6893d44
Sep  2 13:22:52 kernel: x3 : 0000000000000018 x2 : 0000000000021000
Sep  2 13:22:52 kernel: x1 : 0000000000000000 x0 : 0000000000000018
Sep  2 13:22:52 rc_service: service 7065:notify_rc restart_firewall

 

[wuwentao]

You still have that option available?

I'm sorry, it should be :[Enable JFFS custom scripts and configs] and option is [yes]
so it should be my description error, just follow the old mode.

 

[bbunge]

Noticed a crash yesterday. Checking my log, I had 2 crashes logged. But the normal closed source log. Updated TrendMicro’s signature yesterday.Otherwise this firmware has been solid for me, only the occasional 160MHz drop because of radar signals, I think.

Sep  2 13:22:52 kernel: potentially unexpected fatal signal 11.
Sep  2 13:22:52 kernel: CPU: 0 PID: 1254 Comm: asd Tainted: P           O    4.1.52 #2
Sep  2 13:22:52 kernel: Hardware name: Broadcom-v8A (DT)
Sep  2 13:22:52 kernel: task: ffffffc02407f540 ti: ffffffc0280d0000 task.ti: ffffffc0280d0000
Sep  2 13:22:52 kernel: PC is at 0xf69c1b74
Sep  2 13:22:52 kernel: LR is at 0xf6500018
Sep  2 13:22:52 kernel: pc : [<00000000f69c1b74>] lr : [<00000000f6500018>] pstate: 20010010
Sep  2 13:22:52 kernel: sp : 00000000f6893cd8
Sep  2 13:22:52 kernel: x12: 00000000f650001c
Sep  2 13:22:52 kernel: x11: 0000000000000010 x10: 00000000f6500010
Sep  2 13:22:52 kernel: x9 : 00000000f6893e44 x8 : 0000000000000002
Sep  2 13:22:52 kernel: x7 : 00000000f6500010 x6 : 00000000f6501d38
Sep  2 13:22:52 kernel: x5 : 000000000001f2c8 x4 : 00000000f6893d44
Sep  2 13:22:52 kernel: x3 : 0000000000000018 x2 : 0000000000021000
Sep  2 13:22:52 kernel: x1 : 0000000000000000 x0 : 0000000000000018
Sep  2 13:22:52 rc_service: service 7065:notify_rc restart_firewall

I believe the Trend Micro signature file was updated twice in the past week. Now at 2.420. Several days ago I was having issues with the node showing disconnected when running Merlin firmware on router and node. Could be something Trend Micro related?

 

[Ripshod]

I believe the Trend Micro signature file was updated twice in the past week. Now at 2.420. Several days ago I was having issues with the node showing disconnected when running Merlin firmware on router and node. Could be something Trend Micro related?

I have to support this, especially as it was ASD that crashed.

 

[Mogsy]

I believe the Trend Micro signature file was updated twice in the past week. Now at 2.420. Several days ago I was having issues with the node showing disconnected when running Merlin firmware on router and node. Could be something Trend Micro related?

If I get this crash, I always check signature update. Yes it was updated twice.