smart(8) Call for Testing

[Charles Sprickman]

-- 
Charles Sprickman
NetEng/SysAdmin
Bway.net - New York's Best Internet www.bway.net
spork at bway.net - 212.982.9800



> On Mar 29, 2018, at 9:43 AM, Lev Serebryakov <lev at freebsd.org> wrote:
> 
> On 29.03.2018 16:27, Chuck Tuffli wrote:
> 
>>>> Maybe one of the vendors who sells FreeBSD as part of an appliance has shown some interest in this?
>>>> 
>>>> If you’re hardware is well-defined and thus the output is consistent, I could imagine it’s not too difficult to parse this.
>>> smartd is very important part of smartmontools, smartctl is not so.
>>> 
>>> And periodic self-test triggering & check is most important feature of
>>> smartd, IMHO.
>>> 
>>> Modern HDDs are liers in SMART. And only regular self-test discover
>>> real errors on surfaces in my experience.
>>> 
>>> So, tool without support for HDD self-tests is of little usage for
>>> appliances, IMHO.
> 
>> Thank you for the feedback! As I don't have any experience with
>> smartd, can you help me better understand which parts of it are most
>> useful to you? Is it just  periodically triggering the self test or
>> are there other features as well? For example, logging the SMART
>> values, emailing / triggering notifications when certain criteria are
>> met, monitoring the self tests, reading the error logs, etc.?
> 
>  Triggering of short and full self-tests and alerting (via e-mail) when
> test failed.
> 
>  Monitoring of values and alerting is VERY important (number of
> Relocations is main indicator of spinning HDD health and when it raises
> it must be known ASAP), but it could be implemented with simple smart(8)
> utility and some scripting, so this is not problem.

It would be nice if we could grab these values via snmp…  Right now I use
either nrpe or check_by_ssh in nagios to run scripts to parse smartctl output
and it would be weird to have SMART functions in base but not have that
tied to the stock snmpd.  

> But all my dead HDDs were replaced on self-test fail — it is what
> allows me to replace them BEFORE data were lost.

Yep, lots of folks claim the data is useless, but generally I see some signs of
failure before the drive dies, and sometimes those signs are spotted because
smartd is triggering regular self-tests.  And on SSDs, watching the MWI seems 
to work very well - these drives are much smarter (no pun intended) than spinny 
disks.

Charles

> 
> -- 
> // Lev Serebryakov