Operating System - HP-UX
1819988 Members
3475 Online
109608 Solutions
New Discussion юеВ

Re: Temperature exceeded the normal operation threshold.

 
SOLVED
Go to solution
Manuales
Super Advisor

Temperature exceeded the normal operation threshold.

Hi ..
I'd like to know how works /usr/sbin/envd daemon ...

how is detected when the environment is hot and it is sent the message into /var/adm/syslog:

Oct 28 11:02:40 unix3 /usr/sbin/envd[1127]: ***** OVERTEMP_CRIT WARNING *****

Please let me know.
Thanks.
12 REPLIES 12
James R. Ferguson
Acclaimed Contributor
Solution

Re: Temperature exceeded the normal operation threshold.

Hi Manuales:

The manpages offer quite a bit of information (as always);

http://www.docs.hp.com/en/B2355-60105/envd.1M.html

Regards!

...JRF...
Manuales
Super Advisor

Re: Temperature exceeded the normal operation threshold.

ok Mr. James .... thanks for your support in advance ..

ok ..

i'd like to do the following ..
if i want to email a message to my outlook (manuales@patito.com) .. how would do that?

I mean, i have a script located into
/usr/manuales/alert1.sh
alert1.sh contains a line to send a message ...

how can i configure this scirpt into /etc/envd file? i do not understand that ...

what do i have to do to activate the new configuration of the file /etc/envd file withouth shutdown the server?

could you please let me know that?

THANks.


James R. Ferguson
Acclaimed Contributor

Re: Temperature exceeded the normal operation threshold.

Hi (again) Manuales:

There are actually examples of user action in the 'envd' manapges. To specifically answer your question, you would code somethink like:

OVERTEMP_CRIT:y
/usr/bin/mailx -s "SERVER TOO HOT" manuales@xyz.com < /dev/null

That said, the bigger problem is the excessive heat can PERMANENTLY damage your hardware. You should make sure that you have environmental monitoring tools in your computer room that alert you and your site's maintenance staff long, long before you ever receive alerts like this!

Regards!

...JRF...




Manuales
Super Advisor

Re: Temperature exceeded the normal operation threshold.

James, i'm trying to give you your points but system only sends me the following: page can not be display .. i will do it later ..

and ...

if i want system tells me that fans are running right:

did i add the following:

FAN_NORMAL:y
/home/dani/fan_normal.sh




cat /home/dani/fan_normal.sh
- if date = 07:00 then send a message:
/usr/sbin/mail -s "fans right..." ....


is that right?








James R. Ferguson
Acclaimed Contributor

Re: Temperature exceeded the normal operation threshold.

Hi (again) Manuales:

> did i add the following:

FAN_NORMAL:y
/home/dani/fan_normal.sh

Yes, that should work.

However, generally humans become very complacent after reading (or seeing) messages that almost always tell about a condition that can be ignored. You are much more likely to get action (appropriate reaction) when you post only alerts that NEED ATTENTION. The old adage of the "squeaky wheel gets the oil" usually applies :-))

Regards!

...JRF...
Manuales
Super Advisor

Re: Temperature exceeded the normal operation threshold.

thanks James ..

this is only to verify that it is going to work the alert when temperature is higher ..

what do i have to do once i have updated the /etc/envd.conf file ?'
i mean, do i have to start the daemon envd again?

please let me know ..

James R. Ferguson
Acclaimed Contributor

Re: Temperature exceeded the normal operation threshold.

Hi Manuales:

To restart the 'envd' daemon you send your favorite signal --- SIGHUP to the pid of the 'envd' process.

I teasingly say your "favorite" because of another post of yours today.

SIGHUP is the standard signal used with daemons to cause them to re-read their configuration file.

Regards!

...JRF...
Manuales
Super Advisor

Re: Temperature exceeded the normal operation threshold.

yeap, that is my favorite .. i will never its means .. jajajaja...

tell me please:
1.- how do i have to run the signal to the envd daemon? (i'm new on this, really)

2.- how often the envd daemon will run the script that i have configured? each 2 seconds, each 5 minutes, how can i know that?

please let me know :)







James R. Ferguson
Acclaimed Contributor

Re: Temperature exceeded the normal operation threshold.

Hi Manuales:

The safe way to find a process by name is to use the UNIX95 option to enable 'ps' to find a process by name:

# UNIX95= ps -C envd -o pid=

This will return the pid of the 'envd' daemon. Note that "UNIX95=" is followed by whitespace and then the 'ps' command without any semicolon. THis keeps the UNIX95 (XPG4) behavior constained to the command line only.

Once you have the pid do:

# kill -hup

Notice that although we name the hangup signal SIGHUP, we drop the SIG prefix to use it.

I actually don't believe that you are going to see any messages generated *unless* there is a change in the state of the variable you are monitoring. The manpages note that "Whenever an environmental state changes from one level to another...the warning message, if specified, is logged, and the corresponding action is executed once, and only once, per state change." Hence, I think you may have to take on faith that your setup works (OR manually stop a fan; let the alert be generated; and then restart the fan).

Regards!

...JRF...
Manuales
Super Advisor

Re: Temperature exceeded the normal operation threshold.

thanks in advance James ... really i'm learning a lot ..ok ... i have faith but i want to see the alert running !! :)

ok .. i will do that you are mentioned, to stop a fan to verify that ...

i have other 2 questions:

2.- how often the envd daemon will run the script that i have configured? each 2 seconds, each 5 minutes, how can i know that?

3.- what advantages do i have to use UNIX95 command instead of only "ps -fea | grep envd" ...

<:)
James R. Ferguson
Acclaimed Contributor

Re: Temperature exceeded the normal operation threshold.

HI Manuales:

> how often the envd daemon will run the script that i have configured? each 2 seconds, each 5 minutes, how can i know that?

See my answer in the preceding post to this thread; from the manpages: "Whenever an environmental state changes from one level to another (such as from NORMAL to OVERTEMP_CRIT or from FANFAIL_CRIT to FANFAIL_EMERG), the warning message, if specified, is logged, and the corresponding action is executed once, and only once, per state change."

> what advantages do i have to use UNIX95 command instead of only "ps -fea | grep envd" ...

The advantage is that you get an *exact* match of the process basename to the process table. That is, you don't falsely match processes that have the sought after string of characters *somewhere* in their name (basename or command name) or somewhere in their command arguments.

Regards!

...JRF...
Manuales
Super Advisor

Re: Temperature exceeded the normal operation threshold.

James ..

thanks a lot !!! today i learnt how to work with the system when there are problems with high temperature and also i understood the signals .. i have been read the man "command" ....

really, thanks, thanks a lot ...

one my favorite sentence:
"never stop learning ..."

also:
never stop loving
never stop asking
never stop reading
never stop seeing God's creation
never, never ... :)