Bad blocks appeared following shutdown/poweroff

Under what circumstances would turning off the power cause
a disk to develop a pair of adjacent unreadable blocks?
Or do you think they were already there and the reboot just
revealed them?

Background...

We have 4 MAXTOR Atlas 10K V 147GB drives in a JBOD
attached to a SunFire V880 through a Sun Dual Ultra
320 SCSI controller. It has run for half a year with no
disk errors logged by Solaris. A few weeks ago I shut
the system down with:

/usr/sbin/shutdown -g 0 -y -s 5

The V880 powered itself off. I then pulled the 3 power cords
from the power supplies on the back, then turned off the
two power supplies on the JBOD.

Later when power was reapplied (in the opposite order) the system came
up and there were 2 adjacent unreadable bad blocks on one of these
drives. These were eventually forced to remap by:

format -> analyze -> read

on that disk. Since then no other bad blocks have appeared.

Now as I understand it when power is removed from modern hard drives the
heads automatically snap back to some position where they can't damage the data. It looks like here there may have been a bit of contact between the heads and one platter on the way back to this position. The blocks in question live in the middle of an index table from a large Oracle database and I'm guessing that Oracle probably would have noticed
them before this. But maybe not, they could have been latent for a
while and were only revealed when the system was rebooted and Oracle went through some sort of start up verification phase and found them.

Anybody want to rate the relative likelihood of these two scenarios?

Thanks,

David Mathog
.

Relevant Pages

  • Re: Ext4 and the "30 second window of death"
    ... A - Writing data to disk immediately and lose no work at all, ... application uses fsync, right?). ... The problem comes when you have lots of applications open on the ... in order to save power. ...
    (Linux-Kernel)
  • Re: Future Linux on Bistable Storage
    ... One major difference between disk and RAM is the tradeoffs between size, ... resume the system -- except perhaps for I/O initialisation. ... Writing all of RAM to disk burns more power than powering RAM for several ...
    (Linux-Kernel)
  • Re: Spinning down an old disk array
    ... the RD53 started to spin up, but never accelereated to normal speed. ... But for a newer generation SCSI disk, it had come off an old ... Motor and bearing failures are noticable, ... In the case of a disk array, my concern would be with regards to power. ...
    (comp.os.vms)
  • Re: power off disk drives while running
    ... and massive data storage. ... The ability to fully turn off a drive while ... An IDE reset bringing the disk up again -- that does not sound like ... Power down for me means: as if the plug was pulled. ...
    (Linux-Kernel)
  • error using cdrecord
    ... Indicated writing power: 3 ... Disk sub type: High speed Rewritable media ... ATIP start of lead in: -11240 ... Writing time: 14.084s ...
    (comp.unix.solaris)