ASUS RT-AC68U SNMP

[pmsousa]

Hi,

Has anyone accessed the SNMP OID's for CPU Temperature? I'm able to monitor CPU (cores) Load, interface traffic, system menory but can't find the right OID's for the CPU Temp.

Anyone?!?!

Thanks,
Pedro

Firmware: 3.0.0.4.376.49_5

 

[RogerSC]

You may already know this, but if you use Merlin firmware for the RT-AC68U, you get the CPU temperature in the "Tools" page, along with the radio temperatures. It isn't available for using with SNMP software, but you can look at it if you want *smile*.

 

[pmsousa]

You may already know this, but if you use Merlin firmware for the RT-AC68U, you get the CPU temperature in the "Tools" page, along with the radio temperatures. It isn't available for using with SNMP software, but you can look at it if you want *smile*.

Hi RogerSC,
Has you've guessed, I already knew that!

SNMP protocol provides access to sensors like temperature through lm_sensors MIB and I've done it on other systems. When monitoring a large scale network with several other systems being monitored, to look at a particular device page isn't an option. I have a centralized management console and through PRTG monitor, I probe several system with ASUS router included. I've managed to get the values for the ASUS interfaces, ethernet and wireless, CPU (both cores) % usage, RAM (used and available). The ASUS implements MIB's U. C. Davis (2021), ASUS RT-N66U (2623) and NET-SNMP (8072). None of those provide OID's for CPU Temperature...

Lets see if anyone else has done this and knows the OID, if available. Meanwhile, I will try - based on your suggestion - to see the code behind the router's firmware page to find where they get the value from and figure a process to send it to my PRTG probes...

 

[RogerSC]

Sounds like a good plan.

You might try /proc/dmu/temperature. This contains the cpu temperature for most routers that use embedded Linux OS.

 

[pmsousa]

Sounds like a good plan.

You might try /proc/dmu/temperature. This contains the cpu temperature for most routers that use embedded Linux OS.

Hi RogerSC,

I found on the router's .asp pages that the CPU Temperature is obtained by:

curr_coreTmp_cpu = "<% get_cpu_temperature(); %>";

Actually there's a ajax.coretmp.asp with those values:

curr_coreTmp_2_raw = "<% sysinfo("temperature.2"); %>";
curr_coreTmp_2 = (curr_coreTmp_2_raw.indexOf("disabled") > 0 ? 0 : curr_coreTmp_2_raw.replace("&deg;C", ""));
curr_coreTmp_5_raw = "<% sysinfo("temperature.5"); %>";
curr_coreTmp_5 = (curr_coreTmp_5_raw.indexOf("disabled") > 0 ? 0 : curr_coreTmp_5_raw.replace("&deg;C", ""));
curr_coreTmp_cpu = "<% get_cpu_temperature(); %>";

I've managed to read those from my PRTG Probe but got stuck on another issue. The Web Interface only allows for 1 active Admin login, so if the probe is monitoring the page, it needs to be authenticated and ellas I can't login on the web interface!!! Bummer...

I've also built a custom .asp page to get a single value but it's not persistent and when the router reboots, it's gone! Because of the previous issue, I won't explore this solution anymore.

Today, I'll try your method and probably test some kind of scheduled cron task to write a text file with the value and then SCP the file to the probe.

Pedro

 

[pmsousa]

Hi RogerSC,

Houston, we have ignition! I've done it!!!

Using you advice to try /proc/dmu/temperature, I've found that the ASUS Router has that value. From there it was easy to parse the value using grep regex and have it return the expected value for PRTG SSH Probe: "0:value:OK".
I've created my script on JFFS partition and using the *init-start script, symlinked it to the /var folder where PRTG expects it to be.
Now I just have to wait a little longer so I can reboot the router at lunch time and check that everything works has expected.

cpu_probe code:

cat /proc/dmu/temperature | grep -o -e "[0-9][0-9]" | awk '{print "0:"$1":OK"}'

Again, thanks RogerSC for the brainstorming that got me to the final solution.

Pedro

 

[SIdCharming]

Hi pmsousa, I am trying to setup two ASUS routers (model) with MRTG. Currently MRTG is working based on stock values given and I am looking for the interfaces specifically. I am trying the method of watch values coming out to see if I can distinguish them apart. Can you either provide or point me to a good source for the OIDs? (Ps, I've tried MIBwalking with command prompts in Linux, but again, going in somewhat blind. I know enough to be dangerous!) )

 

[RMerlin]

The OIDs are all standard Linux ones. I disabled the Asus-specific OIDs because they were exposing security-sensitive info such as the router login to anyone on the LAN who knew how to use an SMNP client.

 

[SIdCharming]

The OIDs are all standard Linux ones. I disabled the Asus-specific OIDs because they were exposing security-sensitive info such as the router login to anyone on the LAN who knew how to use an SMNP client.

Is that the same case when the router is flashed with DD-WRT and using SNMP?

 

[RMerlin]

Is that the same case when the router is flashed with DD-WRT and using SNMP?

No idea what DD-WRT exposes OID-wise.

 

[SIdCharming]

No idea what DD-WRT exposes OID-wise.

Ah, I will have to investigate. Thanks!

 

[SIdCharming]

The OIDs are all standard Linux ones

Thinking that you are implying OIDs for LINUX vs for the Routers in this statement. I am not monitoring Linux, but monitoring with Linux to a router ASUS RT-N66U in hope of identifying what ports are being measured in MRTG. I wasn't able to test the SNMP that is in the current ROM DD-WRT.