What's new

[Release] Asuswrt-Merlin 384.13 is now available

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

Status
Not open for further replies.
Updated dirty to 384.13 (from 384.9) on both the 86U and 68R, so far so good - fixed some intermittently issues with a few client's WiFi that have been recently updated including a Roku Ultra updated to 9.2, a Samsung Galaxy S7 (after update this month) and a laptop. All seems well and stable, WiFi performance may be down a bit overall but no connection issues, 86U memory usage now at 85% (started at 75%).

.
 
RT-AX88U with V384.13_0 Final was up for 43+ days straight, with over 30 devices attached, downloads, streaming, etc., all working as expected. The GUI eventually became unresponsive, and could not login to the GUI. Tried to SSH into the router, no go. Eventually the router came back online, as it had done a reboot of the router. I believe this is due to a memory leak, as I then cold booted the router, and was up for over 37+ days, everything working flawlessly before the same issue occurred.

Gonna try the Alpha2 V384.14 latest for the AX88U, as the following fix may be related to this issue:

- FIXED: Webui wouldn't notify when running dangerously low on
free nvram (feature was lost at some point in the past)
 
everything working flawlessly before the same issue occurred

AiProtection eats the memory slowly, it's been like this forever. Not 384.13 specific issue, not under RMerlin's control. RT-AX88U has 1024MB RAM and this is why it lasted so long with no reboot needed. RT-AC86U has 512MB RAM and needs a reboot more often. If all TrendMicro components are enabled - about every week. NVRAM is unrelated, issues start when the main memory is about 95% full.
 
  • Like
Reactions: #TY
AiProtection eats the memory slowly, it's been like this forever. Not 384.13 specific issue, not under RMerlin's control. RT-AX88U has 1024MB RAM and this is why it lasted so long with no reboot needed. RT-AC86U has 512MB RAM and needs a reboot more often. If all TrendMicro components are enabled - about every week. NVRAM is unrelated, issues start when the main memory is about 95% full.

I did not experience this with 384.9 and went months without rebooting (and 86U memory usage stable). With 384.13, my 86U has been creeping up, now at 87%.

Yesterday:
86U memory usage now at 85% (started at 75%).

Updated to 384.13 on 10/23/2019, AM
 
Last edited:
I did not experience this with 384.9

Don't remember 384.9, honestly. I don't use any TrendMicro services, but test them on every firmware update to see if some things got fixed. The AiProtection memory leak is still there, not remembering Adaptive QoS user settings on reboot is still there, etc. But RMerlin has no control over TrendMicro components. They will get fixed at some point by TrendMicro/ASUS... in 2023-2025. Have a little more patience. :)
 
I do not have any memory leak wit all TrendMicro services enabled.
Memory usage didn't budge 30+ days...
 
Code:
Oct 26 06:15:49 RT-AC5300 kernel: tdts_core_ioctl_udb_op_prog_ctrl() fail!
Oct 26 06:16:24 RT-AC5300 kernel: sh invoked oom-killer: gfp_mask=0xd0, order=2, oom_adj=0, oom_score_adj=0
Oct 26 06:16:24 RT-AC5300 kernel: [<8005701c>] (unwind_backtrace+0x0/0xf8) from [<800b12d0>] (dump_header.clone.7+0x6c/0x184)
Oct 26 06:16:24 RT-AC5300 kernel: [<800b12d0>] (dump_header.clone.7+0x6c/0x184) from [<800b1590>] (oom_kill_process.clone.9+0x68/0x178)
Oct 26 06:16:24 RT-AC5300 kernel: [<800b1590>] (oom_kill_process.clone.9+0x68/0x178) from [<800b1964>] (out_of_memory+0x154/0x2bc)
Oct 26 06:16:24 RT-AC5300 kernel: [<800b1964>] (out_of_memory+0x154/0x2bc) from [<800b4c84>] (__alloc_pages_nodemask+0x564/0x58c)
Oct 26 06:16:24 RT-AC5300 kernel: [<800b4c84>] (__alloc_pages_nodemask+0x564/0x58c) from [<800b4cbc>] (__get_free_pages+0x10/0xfc)
Oct 26 06:16:24 RT-AC5300 kernel: [<800b4cbc>] (__get_free_pages+0x10/0xfc) from [<80059808>] (get_pgd_slow+0x14/0x158)
Oct 26 06:16:24 RT-AC5300 kernel: [<80059808>] (get_pgd_slow+0x14/0x158) from [<80075f5c>] (mm_init.clone.50+0x9c/0xdc)
Oct 26 06:16:24 RT-AC5300 kernel: [<80075f5c>] (mm_init.clone.50+0x9c/0xdc) from [<800e5710>] (bprm_mm_init+0x10/0x18c)
Oct 26 06:16:24 RT-AC5300 kernel: [<800e5710>] (bprm_mm_init+0x10/0x18c) from [<800e5b7c>] (do_execve+0xcc/0x25c)
Oct 26 06:16:24 RT-AC5300 kernel: [<800e5b7c>] (do_execve+0xcc/0x25c) from [<8005403c>] (sys_execve+0x34/0x54)
Oct 26 06:16:24 RT-AC5300 kernel: [<8005403c>] (sys_execve+0x34/0x54) from [<80050ac0>] (ret_fast_syscall+0x0/0x30)
Oct 26 06:16:24 RT-AC5300 kernel: Mem-info:
Oct 26 06:16:25 RT-AC5300 kernel: Out of memory: Kill process 13116 (dnsmasq) score 16 or sacrifice child
Oct 26 06:16:25 RT-AC5300 kernel: Killed process 13117 (dnsmasq) total-vm:1336kB, anon-rss:0kB, file-rss:0kB
Oct 26 06:22:15 RT-AC5300 kernel: tdts_core_ioctl_udb_op_prog_ctrl() fail!
Oct 26 06:25:27 RT-AC5300 kernel: tdts_core_ioctl_udb_op_prog_ctrl() fail!
Oct 26 06:30:15 RT-AC5300 kernel: tdts_core_ioctl_udb_op_prog_ctrl() fail!
Oct 26 06:32:53 RT-AC5300 kernel: tdts_core_ioctl_udb_op_prog_ctrl() fail!
Oct 26 06:35:19 RT-AC5300 kernel: tdts_core_ioctl_udb_op_prog_ctrl() fail!
Oct 26 06:37:33 RT-AC5300 kernel: tdts_core_ioctl_udb_op_prog_ctrl() fail!
Oct 26 06:37:43 RT-AC5300 pppd[545]: Connect time 7054.4 minutes.
Oct 26 06:40:10 RT-AC5300 kernel: tdts_core_ioctl_udb_op_prog_ctrl() fail!
Oct 26 06:39:55 RT-AC5300 pppd[545]: Sent 899815641 bytes, received 3270013710 bytes.
Oct 26 07:02:21 RT-AC5300 pppd[545]: Connection terminated.
Oct 26 07:21:15 RT-AC5300 pppd[545]: Sent PADT
Oct 26 07:43:17 RT-AC5300 pppd[545]: Modem hangup
The memory leak issue could be HND platform specific, perhaps. @RMerlin may have more information on the subject.

no its there on RT-AC5300 on 384.12 and 384.13. The dump above is representative of what happens after running about a week. Memory starts at 170Mbyte free after reboot around 65% used and then day by day goes to 95% and free ram 25Mb or less and then dumps. I have a 2Gbyte swap - no difference. Never did this under previous firmwares. I am having to schedule reboots every 7 days now. A right pain in the butt.
 
Last edited:
  • Like
Reactions: #TY
RT-AX88U with V384.13_0 Final was up for 43+ days straight, with over 30 devices attached, downloads, streaming, etc., all working as expected. The GUI eventually became unresponsive, and could not login to the GUI. Tried to SSH into the router, no go. Eventually the router came back online, as it had done a reboot of the router. I believe this is due to a memory leak, as I then cold booted the router, and was up for over 37+ days, everything working flawlessly before the same issue occurred.

Gonna try the Alpha2 V384.14 latest for the AX88U, as the following fix may be related to this issue:

- FIXED: Webui wouldn't notify when running dangerously low on
free nvram (feature was lost at some point in the past)




however I don't use AI protect prefer skynet so its turned off.

Memory usage currently sitting at 69%

Will update in a few days but I don't for see myself having this issue never noticed it before even on previous router.
 
no its there on RT-AC5300 on 384.12 and 384.13... day by day goes to 95% and free ram 25Mb or less and then dumps.

Well, I was explaining what happens when the router is at 95% used RAM, but most people didn't believe me.
https://www.snbforums.com/threads/high-ram-usage-rt-ac86u.50070/page-2#post-519508

however I don't use AI protect

With the main suspect disabled there is no memory leak issues.
My router also stays around 60% used memory and can go forever. All TrendMicro services are disabled.
 
Ever since I did a nuclear reset of my router and a clean install of 384.13, Skynet, and Diversion, everything has been running insanely well. I have kept all TrendMicro features off as well. Happy to read that they could have been contributing to all the issues I was having prior to my reset. People should test their software a lot more thoroughly.
 
In certain configurations, the RT-AC86U running 384.13 seems very stable - 61% memory usage after 69 days uptime.

I did a clean install and switched off all the features I don't need:
-clean install
-no AiProtection or other Trend Micro modules
-no Qos
-no AiCloud
-no IPv6

and on the wireless side,
-no Smart Connect
-fixed channels
-no Roaming Assistant
-no Airtime Fairness
-no MIMO
-neither of the Beamforming oprtions enabled
-no Mesh

+OpenVPN server is enabled.

Over time, one of the things mentioned above which are switched off must cause the instability / RAM usage others have noted.

Right now, many seem to consider the Trend Micro modules as suspect #1. But not necessarily - one of the other things mentioned above could also play a part.
 
Can someone advise how to troubleshoot Let's Encrypt certificate generation/update? Since 384.13 it does not work and sits on "Status :Updating..." all the time. I can do get it manually and install, but… it used to work automatically and I would really love to get this functionality back.
 
Can someone advise how to troubleshoot Let's Encrypt certificate generation/update? Since 384.13 it does not work and sits on "Status :Updating..." all the time. I can do get it manually and install, but… it used to work automatically and I would really love to get this functionality back.
I think I read somewhere that it's a big that Asus needs to fix. I simply use the self generated ones. Less hassle and works perfectly.
 
Can someone advise how to troubleshoot Let's Encrypt certificate generation/update? Since 384.13 it does not work and sits on "Status :Updating..." all the time. I can do get it manually and install, but… it used to work automatically and I would really love to get this functionality back.

Nothing can be done about it right now, only Asus guys can fix this since it's part of the closed source code and they're already working on a fix.

https://github.com/RMerl/asuswrt-merlin.ng/issues/373
 
But not necessarily - one of the other things mentioned above could also play a part.

Wireless options do not cause any memory usage issues on RT-AC86U.
Your configuration is the same as mine. It's a router and I use it for routing.
 
up for 43+ days straight, with over 30 devices attached, downloads, streaming, etc., all working as expected. The GUI eventually became unresponsive, and could not login to the GUI. Tried to SSH into the router, no go.

happens to me on other asus router and merlin version,
since the router functions themselves continue as usual,
i just reboot when everyone's sleeping and fogettabouit.

i've gotten in the habit of logging out when i'm done with the gui,
just to see if i can go longer than 2 months without a lockup :rolleyes:
results are still inconclusive as i seem to be rebooting for other
reasons, thus have not gone long enough without interruption.
 
I know of at least one memory leak in the httpd server that was fixed by Asus in 384_81xxx. No idea how severe the leak was, but I can say it was a very arcane one for them to track down...
 
I know of at least one memory leak in the httpd server that was fixed by Asus in 384_81xxx. No idea how severe the leak was, but I can say it was a very arcane one for them to track down...

I'm sitting at 89% this morning and I if I remember correctly, was at 89% when I last checked last night. I will just reboot the if I hit 90-92% and will update to 384.14 once in beta I think....

Adaptive QoS (fq_codel) and AiProtrotection enabled

Now, I am looking into the AiMesh stuff since merlin has it now (sometime after 384.9), trying to decide if I will switch the 68U to current Asus stock and run that as a Node of the 86U (merlin) - essentially switching from a Repeater config.

.
 
Last edited:
Today after about 25 days of flawless operation, I could not log into the GUI on my 5300. This is the second time a hard reboot has been required in this release. Functionally, the router is handling all traffic without a hitch. The logon screen appears, but after credentials are entered, the router never makes it to the first screen?
Does anyone else have a similar experience?
 
Status
Not open for further replies.

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top