[lkml]   [2001]   [Nov]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Journaling pointless with today's hard disks?
On Nov 26, 2001  16:16 -0800, Andre Hedrick wrote:
> On Mon, 26 Nov 2001, Andreas Dilger wrote:
> > What happens if you have a slightly bad power supply? Does it immediately
> > go read only all the time? It would definitely need to be able to
> > recover operations as soon as the power was "normal" again, even if this
> > caused basically "sync" I/O to the disk. Maybe it would be able to
> > report this to the user via SMART, I don't know.
> ATA/SCSI SMART is already DONE!
> To bad most people have not noticed.

Oh, I know SMART is implemented, although I haven't actually seen/used a
tool which takes advantage of it (do you have such a thing?). It would
be nice if there were messages appearing in my syslog (just like the
AIX days) which said "there were 10 temporary read errors at block M on
drive X yesterday" and "1 permanent write error at block M, block remapped
on drive X yesterday", so I would know _before_ my drive craps out
after all of the remapping table is full, or the temporary read errors
become permanent. (I have a special interest in that because my laptop
hard drive sounds like a jet engine at times... ;-).

What I was originally suggesting is that it have a field which can report
to the user that "there were 800 sync/reset operations because of power
drops that were later found not to be power failures". That is what
I was suggesting SMART report in this case (actual power failures are
not interesting). Note also, that this is purely hypothetical, based
on only a vague understanding on what actually happens when the drive
thinks it is losing power, and only ever having seen the hex output
of /proc/ide/hda/smart_{values,thresholds}.

Being able to get a number back from the hard drive that it is performing
poorly (i.e. synchronous I/O + lots of resets) because of a bad power supply
is exactly what SMART was designed to do - predictive failure analysis.

Cheers, Andreas
Andreas Dilger

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:13    [W:0.121 / U:0.176 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site