allow for constant re-sending of warnings

[calestyo]

Hi.

This kinda overlaps with #1017.

Right now, AFAIU, smartd will emit a warning only once per day.
E.g. if my temperature rises over the threshold, I get the warning, but no further one, if it does again (or simply if it stays over the threshold all time).

AFAIU it's the same for other attributes:
E.g. if an reallocated block shows up, I'll get a warning (per day)... but if after that first occurrence (on that day) the number of reallocated blocks increase further, I would not get one until the next day.

Now having one reallocated block, may be not considered that alarming, so on the first occurrence I would perhaps simply say... well... forget about it.
But if more were to come in short time, I'd probably start with backups/replacement immediately.

However, due to the behaviour that such warning comes only once a day,... I wouldn't even notice the further incidents.

For that reason, I think it would be nice if one could set the interval in which new warnings are emitted. Like e.g. "on every case"... or "every 10 mins".

Thanks,
Chris.

[Christian Franke]

Right now, AFAIU, smartd will emit a warning only once per day.
E.g. if my temperature rises over the threshold, I get the warning, but no further one, if it does again (or simply if it stays over the threshold all time).

Quote from smartd.conf man page:
The warning email counter is reset if the temperature dropped below INFO or CRIT-5 if INFO is not specified.

Then a warning email will be sent immediately when the over threshold condition is detected again. If the condition persists, a reminder email will be sent once per day.

AFAIU it's the same for other attributes:
E.g. if an reallocated block shows up, I'll get a warning (per day)... but if after that first occurrence (on that day) the number of reallocated blocks increase further, I would not get one until the next day.

A simple enhancement would be a -M always directive to sent warning emails always when a LOG_CRIT entry is written to syslog.

Events ("X_current > X_previous") should possibly be handled differently from Conditions ("X > LIMIT").

[Christian Franke]

Ticket #1017 has been marked as a duplicate of this ticket.

[Christian Franke]

See also ​GH issues/153.

[Christian Franke]

Added -M always directive in r5428.