[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: fsck on boot...revisited



On 26/07/13 17:53, Ralf Mardorf wrote:
On Fri, 2013-07-26 at 11:54 -0400, Stefan Monnier wrote:
The system in question is running from an SSD, which I assume changes your
assumptions quite a bit. With a traditional HDDs, the loss of power
causes a head crash, etc which does in turn lessen the life of the drive.

Actually, I fail to see why a power outage would have any negative
effect on an HDD (e.g. why would the "head crash"?).

In the old days the heads were not driven back, but for around more than
20 years the heads will park by the momentum of the turns, when there
should be a blackout. IIRC if the drives didn't turn fast enough, the
heads could crash. Today a power outage definitively doesn't cause a
crash and AFAIR even my old 42 MB SCSI survived power outages.

True. Back in those days the heads on some disks had to be sent a "park" command before powering off and what that did was to move the head to an reserved area of the disk before landing. Modern drives park the heads off the disks completely.

I do remember so old 500MB mainframe disk units (14 inch 10 platter or so, 4 per unit) that had a slight fault. The normal procedure was to press the "offline" button, which retracted the heads from the disk, then press the "power" button to turn them off. Due to a glitch in the circuitry this would sometimes result in a surge in the head actuator coil and the heads would move back across the disks just as they stopped spinning, resulting in major damage to the entire unit. Later a small button was fitted which disconnected the coil and we had to hold that down until the disk had stopped spinning.


--
Dom


Reply to: