Custom Query (1566 matches)

Server: Debian squeeze smartmontools 5.39.1+svn3124-2

Hello,

The temperature I get from the log file is very different from the one I get with the smartctl command:

cat /var/log/syslog | grep Celsius

May  7 07:00:19 interne smartd[17990]: Device: /dev/sda [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 119 to 118
May  7 07:00:19 interne smartd[17990]: Device: /dev/sdc [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 118 to 117
May  7 07:30:18 interne smartd[17990]: Device: /dev/sdc [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 117 to 118
May  7 08:00:18 interne smartd[17990]: Device: /dev/sdb [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 32 to 33
May  7 08:00:19 interne smartd[17990]: Device: /dev/sdc [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 118 to 117
May  7 08:30:19 interne smartd[17990]: Device: /dev/sdb [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 33 to 32
May  7 08:30:20 interne smartd[17990]: Device: /dev/sdd [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 121 to 120
May  7 10:00:20 interne smartd[17990]: Device: /dev/sdc [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 117 to 118
May  7 11:00:19 interne smartd[17990]: Device: /dev/sdb [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 32 to 33
May  7 11:00:19 interne smartd[17990]: Device: /dev/sdd [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 120 to 121
...

And when I use smartctl:

smartctl -x /dev/sda

Current Temperature:                    32 Celsius

Index    Estimated Time   Temperature Celsius
  43    2013-05-07 12:59    32  *************
 ...    ..( 42 skipped).    ..  *************
  86    2013-05-07 13:42    32  *************
  87    2013-05-07 13:43    31  ************
 ...    ..( 15 skipped).    ..  ************
 103    2013-05-07 13:59    31  ************
 104    2013-05-07 14:00    32  *************
 ...    ..(122 skipped).    ..  *************
 227    2013-05-07 16:03    32  *************
 228    2013-05-07 16:04    33  **************
 ...    ..( 60 skipped).    ..  **************
 289    2013-05-07 17:05    33  **************
 290    2013-05-07 17:06    32  *************
 ...    ..(111 skipped).    ..  *************
 402    2013-05-07 18:58    32  *************
 403    2013-05-07 18:59    33  **************
 ...    ..( 58 skipped).    ..  **************
 462    2013-05-07 19:58    33  **************
 463    2013-05-07 19:59    32  *************
 ...    ..(  7 skipped).    ..  *************
 471    2013-05-07 20:07    32  *************
 472    2013-05-07 20:08    31  ************
 ...    ..( 25 skipped).    ..  ************
  20    2013-05-07 20:34    31  ************
  21    2013-05-07 20:35    32  *************
 ...    ..( 20 skipped).    ..  *************
  42    2013-05-07 20:56    32  *************

(same values for sdb, sdc & sdd)

Any idea where that comes from?

119° Celsius.. Imagine my face when I saw that :)

While trying to evaluate the numbers smartd write to attrlog files, I noticed that for SCSI disks the attributes have names like "temperature", while for ATA disks there are no attribute names, i.e. the attribute is specified by number (like 194). I'd prefer (ore or less) short and stable attribute names if known.

For example a line in an ATA attrlog looks like: 2015-12-18 12:10:28; 1;63;38776360; 3;98;0; 4;99;1185; 5;100;0; 7;84;290151632; 9;68;28632; 10;100;0; 12;99;1191; 194;41;55834574889; 195;63;38776360; 197;100;0; 198;100;0; 199;200;0; 200;100;0; 202;100;0;

A line from a SCSI attrlog looks like: 2016-03-08 15:05:07; read-corr-by-ecc-fast;0; read-corr-by-ecc-delayed;26075; read-corr-by-retry;0; read-total-err-corrected;26075; read-corr-algorithm-invocations;0; read-gb-processed;337052.962; read-total-unc-errors;0; write-corr-by-ecc-fast;0; write-corr-by-ecc-delayed;48719; write-corr-by-retry;0; write-total-err-corrected;48719; write-corr-algorithm-invocations;0; write-gb-processed;1481.749; write-total-unc-errors;0; verify-corr-by-ecc-fast;0; verify-corr-by-ecc-delayed;0; verify-corr-by-retry;0; verify-total-err-corrected;0; verify-corr-algorithm-invocations;0; verify-gb-processed;0.000; verify-total-unc-errors;0; non-medium-errors;366; temperature;28;

Hi.

The minimum check interval for smartd seems to be 10 minutes and the default seems to be 30 minutes.

Is there any reason for that limitation (like more often checks having influence on the drive health)?

If not than I think allowing smaller minimums (perhaps 60 or 120s?) and a smaller default (5-10 mins?) would have several benefits:

1) Removable devices, are often only attached for short times,... so with the default of 30 mins, they may not get checked at all for a long time. (Though the proper solution for this is actually #1016).

2) I've seen drives starting to die pretty sudden without showing any big and steady degradation of SMART values before... but rather a very abrupt one. So for such cases it may be helpful, to check the drives more often, in order to notice this.

3) Making smartd usable as a hddtemp replacement (see #1017). Temperatures can increase quite fast,... to make smartd usable as a temperature monitor, it would be nice if much smaller check intervals were possible.

Cheers, Chris.