[Beta] Asuswrt-Merlin 384.10 Beta is now available

[dugaduga]

Could dirty flashing or importing saved asuswrt merlin cfg settings from the previous versions over this new openvpn 2.4.7 be causing this?

 

[L&LD]

Could dirty flashing or importing saved asuswrt merlin cfg settings from the previous versions over this new openvpn 2.4.7 be causing this?

I was just going to ask you this! I think you hit the nail on the head.

Follow the M&M Config guide and the other links in my signature below to put your router into a good/known and stable state. Do not use your saved config files.

In this order; I would do a WPS NVRAM erase cycle, reset to factory defaults and also format the jffs on next reboot followed by rebooting the router 3 times and waiting 5 to 10 minutes between reboots before manually and minimally configuring the router to secure it and connect to your ISP.

 

[PeterV]

Today I'm first time faces with a seg-fault, i.e. my AC86U rebooted due to kernel/hw error on 384.10 Beta 2( it was a dirty upgrade yesterday morning).
It was happens on today evening, after one day and 10 hours uptime.

May  5 08:05:05 kernel: _ Reboot message ... _______________________________________________________
May  5 08:05:05 kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000000pgd = ffffffc00079e000[00000000] *pgd=000000001e806003, *pud=000000001e806003, *pmd=000000001e807003, *pte=
0060000081001707Internal error: Oops: 96000006 [#1] PREEMPT SMPModules linked in: ebt_mark init_addr(          (null) -           (null)), core_addr(ffffffbffc156000 - ffffffbffc156300) ebtable_nat ...
...
May  5 08:05:07 kernel: al exception in interruptCPU0: stoppingCPU: 0 PID: 210 Comm: bcmsw_rx Tainted: P      D    O    4.1.27 #2Hardware name: Broadcom-v8A (DT)Call trace:[<ffffffc0000876d8>] dump_backtrace
+0x0/0x150[<ffffffc00008783c>] show_stack+0x14/0x20[<ffffffc00050851c>] dump_stack+0x90/0xb0[<ffffffc00008e730>] handle_IPI+0x190/0x1a0[<ffffffc000080c70>] gic_handle_irq+0x88/0x90Exception stack(0xffffffc01
7adf620 to 0xffffffc017adf740)f620: 14768540 ffffffc0 fc2ed268 ffffffbf 17adf760 ffffffc0 005
May  5 08:05:07 kernel: 0001 00000000f720: 17adfd30 ffffffc0 17adf760 ffffffc0 fc2c30fc ffffffbf 17adf760 ffffffc0
May  5 08:05:07 kernel: ____________________________________________________________________________

It's look like some how related to bcmsw_rx. Bad signal for me. On weekend I'm going to do all sanitize procedures(reset, clear nvram, etc)

 

[RMerlin]

Check how much free nvram you have left, or your might just recreate your original problem. Your router only has 64KB total.

Sent from my Nexus 5X using Tapatalk

 

[L&LD]

Check how much free nvram you have left, or your might just recreate your original problem. Your router only has 64KB total.

Sent from my Nexus 5X using Tapatalk

Didn't realize that the RT-AC86U has only 64KB NVRAM! Any way to increase it?

 

[PeterV]

Check how much free nvram you have left, or your might just recreate your original problem. Your router only has 64KB total.

Internal Storage
NVRAM usage 66703 / 131072 bytes
JFFS 3.11 / 48.00 MB
How can I monitor it via cli ?

 

[Kingp1n]

Just installed beta 2 and all seems fine on RT-AC86U.

It is normal for my ram to stay at 90-95% while using the following amtm with diversion, pixelserv, skynet & adaptive QoS enabled with FreshJR script?

Update: this was the same on previous FW btw...

 

[skeal]

Just installed beta 2 and all seems fine on RT-AC86U.

It is normal for my ram to stay at 90-95% while using the following amtm with diversion, pixelserv, skynet & adaptive QoS enabled with FreshJR script?

Update: this was the same on previous FW btw...

Yes this is normal, I would deploy a swap file from the amtm menu for best performance.

 

[Gar]

Yes this is normal, I would deploy a swap file from the amtm menu for best performance.

How big a swap file?

 

[skeal]

How big a swap file?

Make one the same size as your ram capacity. Mine is quite big at 1gig on my AX88U but my AC3100 is 512meg.

 

[Deleted member 62525]

Running latest beta for 1d 5hrs with Skynet, Diversion, Stubby and OpenVPN. No issues to report. I can see that RAM usage is lower.

 

[RMerlin]

Any way to increase it?

Not from me.

How can I monitor it via cli ?

nvram show

 

[L&LD]

How big a swap file?

On my amtm + scripts USB drive, I create the maximum it allows (2GB) on all routers I put it on.

 

[Deleted member 62525]

Just installed beta 2 and all seems fine on RT-AC86U.

It is normal for my ram to stay at 90-95% while using the following amtm with diversion, pixelserv, skynet & adaptive QoS enabled with FreshJR script?

Update: this was the same on previous FW btw...

I have the same setup. No FreshJR but in addition running OpenVPN client and I am closer to 80%

 

[dugaduga]

@RMerlin you are right!

NVRAM usage 65554 / 65536 bytes; talk about magic, did you do this merlin?

How do I clean this up? Second time this has happened after factory default, it would seem...

 

[L&LD]

@RMerlin you are right!

NVRAM usage 65554 / 65536 bytes

How do I clean this up? Second time this has happened after factory default, it would seem...

What are you running on your router as features/options/scripts?

Do you do a full factory reset to defaults? Including the WPS NVRAM erase, formatting the jffs on next reboot and installing any amtm + scripts 'clean'.

 

[dugaduga]

Cleared NVRam Cache, formatted jffs,

No other scripts at all, no openvpn clients,
NVRAM usage 52000 / 65536 bytes
With one openvpn client+cert (no other scripts)
NVRAM usage around ~57000 / 65536 bytes
With two openvpn clients+certs: (no other scripts)
NVRAM usage 62604 / 65536 bytes

That is a roughly ~5000 kb increase per opvn, exact size as the openvpn certificate authority.
Cache does not go down after disabling the second client, even after rebooting.
After deleting the certificate and rebooting Nvram showed:
NVRAM usage 62587 / 65536 bytes
Update: After formatting jffs again, NVRAM usage 62607 / 65536 bytes
I suspect trying more than 2 clients/ certs went over the nvram limits previously.
I was not able to to start a second client even within the NVRam limit.

 

[dugaduga]

Fresh boot of third NVram Clear:
NVRAM usage 51320 / 65536 bytes
After Loading cleanly flashed fully configured 384.10_beta2 cfg (includes 2 opnvpn clients confiugred, certs were already previously cleared by jffs format)
62337 / 65536 bytes
bear in mind i saved the cfg when certs were installed

 

[MarkRH]

In the Traffic Classification Tab of the QoS area, when hovering over any label on the right-side, the tooltip always says undefined. I don't have QoS enabled so that might be it if it's supposed to show something. The browser console only has this warning on that page:

The script from “http://router.asus.com/help_content.js” was loaded even though its MIME type (“text/html”) is not a valid JavaScript MIME type.[Learn More] QoS_Stats.asp

 

[dugaduga]

Before adding 3rd client
NVRAM usage 62315 / 65536 bytes
After adding 3rd client (no cert yet)
NVRAM usage 62318 / 65536 bytes
After adding cert:
65534 / 65536 bytes

cert nvcache its passed on with cfg files even with no certs in jffs, these cache certs are not visible in the openvpn configuration gui or jffs. that aint going back down until i clear the nvcache with wps; unless you know of some command or openvpn directive to help that process to eliminate old unused certs from cache, would be nice to see compression here.