RB960PGS (hEX POE) crashing

Hi All.
Seeking some support with my router. It crashes completely any time I try to do large file transfers on my LAN (between either my PC and NAS, or between two NAS). I want some guidance on how to troubleshoot it, which menus etc might provide the best insights. I didnt find the logs to be meaningful but they might need to be configured in some way. If additional info is needed, export of config etc please let me know.

In terms of supporting info, I have the below.

Physical Connections:
Port 1 - ISP modem (internet is via PPoE)
Port 2 - Unifi AP (PoE Out)
Port 3 - Synology DS1821+ NAS
Port 4 - CCTV NVR
Port 5 - Netgear 5 port GbE switch
This switch has my unRAID NAS connected and a couple of my HiFi/AV amp, streamer etc

Model 960PGS
Firmware Type qca9550L
Factory Firmware 6.42.10
Current Firmware 7.3.1
Upgrade Firmware 7.3.1
Name voltage
Value 48.8
Type V
Name temperature
Value 74
Type C

Thanks for any help.

What is missing is how is configured and… what power source you use…

OK the power supply is 48V 2A.

Sorry but what is the best way to export the config (I use the “webfig” menu). Im trying to use the make supout option but it seems to stall at 98%. Will keep waiting.

I copied this info during a crash just now -
Uptime 00:50:26
Free Memory 87.0 MiB
Total Memory 128.0 MiB
CPU MIPS 74Kc V5.0
CPU Count 1
CPU Frequency 800 MHz
CPU Load 6 %
Free HDD Space 3976 KiB
Total HDD Size 16.0 MiB
Sector Writes Since Reboot 723
Total Sector Writes 2 877 166
Bad Blocks 0.0 %

opena terminal, and do export command, then after removing serial number and other sensitive data, put the results on forum.

I hope I did it right. I edited a couple of bits of personally identifiable info out of it. It doesnt seem exhaustive I have to say, I was a little surprised to my interfaces all at 100M, expected at least the ports connected to the Synology and the Switch to be Gigabit. I also have the supout file if its helpful.

Anyway, hope this is of some use to diagnose it.
config.rsc (4.58 KB)

Is 48V / 2A by PoE or by jack?

some bit of your config are present because your device is not netinstalled with 7.3.1 or wiped between 6.3x to 6.4x upgrade.
Some values have new defaults and is why keeping old 6.3x settings are reported differents.
speed=100Mbps alone not mean that the port work on 100Mbps, but only when auto-negotiation is disabled.

Paste this on terminal, do not forget starting { and ending } !!! and the item surely not needed for you are set to default 7.3.1 values.

{
/interface bridge
set bridge protocol-mode=none
/interface bridge port
set [find] ingress-filtering=yes
/interface ethernet
set [ find default-name=ether1 ] speed=1Gbps
set [ find default-name=ether2 ] speed=1Gbps
set [ find default-name=ether3 ] speed=1Gbps
set [ find default-name=ether4 ] speed=1Gbps
set [ find default-name=ether5 ] speed=1Gbps
set [ find default-name=sfp1 ] advertise=10M-half,10M-full,100M-half,100M-full
/user group
set full name=full policy=local,telnet,ssh,ftp,reboot,read,write,policy,test,winbox,password,web,sniff,sensitive,api,romon,rest-api,!dude
/ip settings
set max-neighbor-entries=4096
/ipv6 settings
set max-neighbor-entries=2048
/ip traffic-flow
set interfaces=all
/interface ovpn-server server
set auth=sha1,md5,sha256,sha512
/interface lte apn
set [ find default=yes ] ip-type=auto use-network-apn=yes
/system logging
set 3 action=echo
add topics=critical
}

Only check if your ether2 mac address is 74:4D:28:29:8B:07, if not, copy that MAC address and put on bridge admin mac address

THank you for your assistance.

The PoE is via the DC input not via PoE In.

You are right regarding the port settings, if I check each interface I can see it has Autonegotiated to 1Gbps.

REgarding the config you have provided, is that to address the issue I reported re the router crashing, or just to fix some other settings?

I just recheck and add other parts on previous script.
This do not fix any problem, because I do not see any problems inside,
but if the problem is something that is not set to default,
just paste that script on terminal and useless settings that you do not use and are parts of old configuration
are set to default values, aligned to 7.3.1 version

Ok well I did try that, via SSH as I've just learnt the unique way pasting works (or doesnt work) with this Terminal.

It returned an error:
input does not match any value of policy

Obviously my prioroity is working out why transferring files brings down my entire network, the router appears to reset at least from the logs. But theyre not very verbose so guessing a bit.

fixed the script, removed quotes, tikiapp no longer exist, and replaced with rest-api
please retry

OK I ran that and it completed without error. See attached.

Any thoughts on the crashing? It's still happening and Im trying to migrate off unRAID to Synology, ~27TB to do so this is proving a nightmare. I'll probably just replace the unit with something else but I liked the fact it is a Router with PoE so I didn't have to bother with adding a PoE switch to my setup.
config1.rsc (3.84 KB)

I can not see any config error, try to see % cpu and memory used, if during transfer go to 100% or memory is depleted

Yeah neither of those peak unusually.

Is there any way to enable some decent logging with this OS? It seems to log only very basic events.

The big problem of default logging setup is that logs go into memory only and reboot wipes it. Even if problem causing reboot does emit a log entry it gets lost. So connect an USB flash disk to your device and configure logging to write logs to disk as well.

I specifically checked five of my own devices with 3 POE consumers connected to them. the temperature ranges from 58 to 63. You have 74 is a lot, when the load increases, the protection trips and the device reboots.
Solve the cooling problem

Ohhh, nice catch!!!

Thanks for taking the time to check that out. I am working on getting the temp in the cabinet down.

Its strange that the router is working so hard with just 1 x POE device connected, and is so much hotter than the two other things in the cabinet - Synology NAS 51deg and Uniview NVR 43deg. Ambient temp is 24deg. In fact the router is at 76deg right now and Im not doing anything to create excessive load. Maybe its faulty, will try to cool it actively and test.

Thanks again.

Show me a photo of the installation site, maybe we can tell you something.
Try to lower the CPU frequency to 720, it may lower the temperature a little.

P=I²R, and R ≠ 0. The more current through the PoE line, the more power dissipated, so the more heat generated.

Would you post the output of my revised PoE status script? It may prove enlightening.

The more current through the PoE line, the more power dissipated, so the more heat generated.

According to specs, Max out per port output (input 30-57 V) 450 mA
So, i don’t think the heat is caused by that…