On Tue, Jun 18, 2013 at 05:18:25PM +0100, Stan Hoeppner wrote: > On 6/18/2013 10:38 AM, Karl E. Jorgensen wrote: > > > But... Since there are no clear suspects, paranoia dictates a run of > > fsck on the affected file systems. Just in case. At least it is a > > harmless check if you can afford the downtime while the file systems > > are unmounted. > > If the PSU is headed South this may not be harmless. An fsck is going > to generate serious head movement while seeking the metadata inodes. > Head movement increases power draw, sometimes double that of the drive > at idle spin. If errors are found and corrections being written when > the system locks up, due to voltage drop caused by increased power draw > by the drive, then you may have a more corrupted filesystem, not less. > > Find the cause of the lockup and fix it, before performing a destructive > repair of a filesystem. If a system is acting flaky the last thing > anyone should do is a destructive repair, as it just might be that.
Ah yes. That is a very good point. Thanks for highlighting that. In other words, it should be treated as an unexploded hand grenade. Carefully. Perhaps an fsck with -n (to prevent actual repair) would be safer? depends on the filesystem in use... -- Karl E. Jorgensen -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130618172329.GD15076@hawking
