[CentOS] Re: SMART Prefailure Attribute messages
Scott Silva
ssilva at sgvwater.com
Wed Oct 17 15:34:51 UTC 2007
on 10/17/2007 6:59 AM Reuben D. Budiardja spake the following:
> Hello,
> My machine is running Software RAID 5 on /dev/hde1, /dev/hdg1,/dev/hdi1. I
> noticed on the log that suddenly we are getting messages such as :
>
> Oct 14 01:27:16 localhost smartd[4801]: Device: /dev/hde, SMART Prefailure
> Attribute: 8 Seek_Time_Performance changed from 247 to 246
> Oct 14 01:57:15 localhost smartd[4801]: Device: /dev/hde, SMART Prefailure
> Attribute: 8 Seek_Time_Performance changed from 246 to 245
> Oct 14 02:27:15 localhost smartd[4801]: Device: /dev/hde, SMART Prefailure
> Attribute: 8 Seek_Time_Performance changed from 245 to 246
> Oct 14 02:27:15 localhost smartd[4801]: Device: /dev/hde, SMART Usage
> Attribute: 195 Hardware_ECC_Recovered changed from 253 to 252
> Oct 14 02:57:15 localhost smartd[4801]: Device: /dev/hde, SMART Prefailure
> Attribute: 8 Seek_Time_Performance changed from 246 to 245
> Oct 14 02:57:15 localhost smartd[4801]: Device: /dev/hde, SMART Usage
> Attribute: 195 Hardware_ECC_Recovered changed from 252 to 253
>
> Similar messages for /dev/hdi and /dev/hdg also, however it seems
> that /dev/hde has the most and have been occuring the longest.
>
> [root at localhost log]# grep SMART messages messages.1 | grep hde | wc
> 178 3023 25944
> [root at localhost log]# grep SMART messages messages.1 | grep hdg | wc
> 93 1578 13583
> [root at localhost log]# grep SMART messages messages.1 | grep hdi | wc
> 95 1612 13865
>
> This did not happen a week ago:
> [root at pathfinder log]# grep SMART messages.2
>
> Checking /proc/mdstat indicates that all the drives are active i.e. RAID is
> not running in degraded mode.
>
> So, is it that suddenly my drives are failing ?
> Any comment or suggestion is greatly appreciated.
>
That could be possible, or your power supply is slightly underpowered or
losing output.
--
MailScanner is like deodorant...
You hope everybody uses it, and
you notice quickly if they don't!!!!
More information about the CentOS
mailing list