rp3410 automatic down problem

[newwutao]

The server automatic down with high frequence.
After the server down,the PS command out is below:
PS
System Power state: Off
Temperature : Low OverTemp


Power supplies | Fan
# State Type | States
-----------------------------------------------------------
0 Normal Type 0 | Normal
1 Normal Type 0 | Normal
2 - - | Normal
3 - - | Normal
4 - - | Normal
5 - - | Normal
6 - -
7 - - | Normal
what maybe the problem? Is there a hardware problem?
please you help me

[Patrick Wallek]

What is the temperature in your room? It sounds as if your ambient room temperature is too high, which is causing your server to shut down automatically to prevent serious damage.

[Yogeeraj_1]

Hi,

definitely a temperature problem detected by the gsp. Did you detect any abnormal condition on the server room infrastructure?

Also, verify that the air flow is not being affected.


revert
kind regards
yogeeraj

[Torsten.]

Have a look into the MP logs and the OLDsyslog.log file.
Are there any overtemp / overtemp shutdown messages? I'm quite sure you will find some ...

[newwutao]

There is no message about High Temperature in the syslog file and OLDsyslog file.
There is no message in shutdownlog also.
Is there another file I can check?

[newwutao]

I find following configuration in the envd.conf file:

OVERTEMP_CRIT:y

OVERTEMP_EMERG:y
/usr/sbin/reboot -qh

FANFAIL_CRIT:y

FANFAIL_EMERG:y
/usr/sbin/reboot -qh

Because I known the temperature is high from ps command's output, I want to ask if there is another way to check the server down automaicly for reason high temperature.
thanks all.
please give me some advices!

[Matti_Kurkela]

Note that your system seems to be indicating "Low OverTemp" even while system power is *off*.

With that much ambient heat, powering up the server is likely to heat it even more, maybe so much that the low-level overtemperature alert becomes a high-level alert.

If that happens, envd will cause the system to power off _very_ abruptly (reboot -qh) to prevent damage. As this method of halting the system will stop things like syslogd too, there may be no record of the event in the logs of the operating system.

Check the logs of the Management Processor using the SL command. Select E for System Events, then T to see the logs in text mode, then just keep pressing Enter to see more and more log events.

If this is the only server that shows temperature alarms, ensure that something is not blocking the airflow to the server: for example, a loose plastic bag can become stuck against the air intake.

MK