kernel errors and unknown SMART properties

Hi!
My issue is a bit complex. Please bear with me.
I wrote a lenghty article on the subject a year ago:
http://yalis.fr/cms/index.php/post/2013/12/14/“If-it-ain’t-broken,-don’t-fix-it”…-or-should-you

At the time, I decided to just let things be, since I had no clue to a 
solution.
Today, I decided to try some new tests. I ran smartctl with -d marvell, 
ata, scsi, sat, sat,12, sat,16, with -i, -a, -x, diff'ed the results… 
all the time while running tail -f /var/log/syslog | grep kernel.

First, I could see that probably ALL of my kernel errors stem from Smart 
actions.

But all these tests produced loads of Unknown_Attribute. _Until I ran_:
diff <(smartctl -d sat,12 -x /dev/sda) <(smartctl -d sat,16 -x /dev/sda)

And there, for the first time in a very long time, I could see:
>   1 Raw_Read_Error_Rate     POSR-K   200   200   ---    -    0
>   3 Spin_Up_Time            POS--K   229   176   ---    -    3533
>   4 Start_Stop_Count        -O--CK   100   100   ---    -    568
>   5 Reallocated_Sector_Ct   PO--CK   200   200   ---    -    0
>   7 Seek_Error_Rate         -OSR-K   200   200   ---    -    0
>   9 Power_On_Hours          -O--CK   088   088   ---    -    9333
>  10 Spin_Retry_Count        -O--CK   100   100   ---    -    0
>  11 Calibration_Retry_Count -O--CK   100   253   ---    -    0
>  12 Power_Cycle_Count       -O--CK   100   100   ---    -    36
> 192 Power-Off_Retract_Count -O--CK   200   200   ---    -    24
> 193 Load_Cycle_Count        -O--CK   200   200   ---    -    543
> 194 Temperature_Celsius     -O---K   118   110   ---    -    32
> 196 Reallocated_Event_Count -O--CK   200   200   ---    -    0
> 197 Current_Pending_Sector  -O--CK   200   200   ---    -    0
> 198 Offline_Uncorrectable   ----CK   200   200   ---    -    0
> 199 UDMA_CRC_Error_Count    -O--CK   200   001   ---    -    819

So I thought: Oh!, so sat,16 is the right value, and I forgot to specify 
",16" before! Well, except that this is the default value, and sure 
enough, a subsequent run of smartctl -d sat,16 -x /dev/sda just gave me 
Unknown_Attribute. Bummer.

I tried to run this last command again and again, to no avail: only 
Unknown_Attribute :-(

_But_ I found that running with -d sat,12, then -d sat,16, then -d 
sat,12, and so on, would sometimes give me the right output. Only, this 
right output would sometimes occur while running with -d sat,16, and 
sometimes occur while running with -d sat,12! So weird.

Oh, and did I mention, that while it is so hard to get a useful result 
with smartctl, smartd seems to run mostly happily, since I am notified 
every day of a bunch of changes in the temperature property from 118 to 
117, and back to 118, and so on… and a couple of other properties too. 
I'm lost, I don't get the logic of any of this.

I don't want to make this mail any longer than it already is, so I'll 
stop here. But please advise me in kernel configuration, smart 
configuration, anything. Ask questions! About anything! This can be a 
long-winded year-long exploration if need be.
Thanks,

Yves.

------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
_______________________________________________
Smartmontools-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/smartmontools-support