SOLVED SMART -- Emails Report sended just once

[nexus41]

Hello!

I'm a beginner and I'm doing some test with FreeNAS-9.2.1.7.
In order to test the SMART alert email reporting functionality I set "Services > SMART" in this way:

Check Interval: 1
Power mode: Never - Check the device regardless of its power mode
Difference: 0
Informational: 0
Critical: 15
Email to report: my-email@gmail.com

Doing this and applying the changes I immediately receive the alert emails about the detection of a SMART error (Temperature) and I can see in the console a message like this:

Device: /dev/ada0, Temperature 34 Celsius reached critical limit of 15 Celsius (Min/Max 34/34)

After 1 min the console messages are correctly updated, but no more emails are sended to my address. In other word the console messages are continuously updated as expected, but emails are sended just the first time, at the moment that I apply the changes or if I restart the SMART service.

Is this normal or I'm doing something wrong? How to ensure that the emails are sended to my address?

TIA

 

[DrKK]

The email, I believe, is sent once, at the time the device goes into a warnable state. If you let the device come out of the warnable state (i.e., the temperature goes back to the designated zone), then it will send an email again when/if it re-enters the warnable state.

That's my understanding. You get 1 email on each ingress into a warnable state.

So I believe it is "working as intended".

 

[titan_rw]

Yes, because the default with smartd is one email per alert trigger. As DrKK says, if the alert 'untriggers', you'll get another warning when it triggers again.

You may change the default by using the -M option in the 'smart extra options' on a per disk basis. I use "-M daily".

I don't see the need to have smart check every minute. Seems a bit excessive. The default is 30, and seems quite reasonable.

 

[cyberjock]

Once per minute! WOW! To be honest, 15 minutes is about the most conservative I'd go. Querying SMART does actually require some hard drive I/O because the hard drives logs and such are actually on the disk platters. Personally I've done some nasty things with some of my servers (aka no fans on purpose and such) and found that 30 minutes is still pretty reasonable.

 

[nexus41]

Thank you very much for your clarifications!
Now the behaviour of the SMART email alert system is clear.

I adopted the settings above (smart check every minute and critical temperature to 15 °C) just with the purpose to test the email alert system.
My settings are now as follow:

Check Interval: 30
Power mode: Never - Check the device regardless of its power mode
Difference: 0
Informational: 0
Critical: 40
Email to report: my-email@gmail.com