Sorry, for my report, as it was badly misleading. That the fs should be checked is out of question, if the datestamp is in the future. But that was not the issue I had in mind. The messages by fsck (ext3) was like this:
"Superblock mount time in Future -> Check forced (...) Disk gone unchecked since 46905 days. Check forced" The second one is wrong, considering, that fsck already determined that something happened to the filesystem in the future. It should not be printed, or if printed, not saying that it was last checked 125 years ago... On Wed, 2007-10-10 at 08:15 -0400, Theodore Tso wrote: > On Tue, Oct 09, 2007 at 09:08:16PM +0200, Tobias Frost wrote: > > Package: e2fsprogs > > Version: 1.40.2-1 > > Severity: minor > > File: /sbin/fsck > > > > My mainboard has a little problem: Rarely, it manipulates its clock to a > > date in the future. Also it did today: It set the time to Oct 16. > > > > So I rebooted to fix the date in the bios... > > After the change, I booted as normal but the filesystem-check fired: > > > > It telled me, that the filesystem has been unchecked for approx. 46000 > > days and then starting the check. > > > > I think, this could be an unsigned-signed issue.... > > Well, all of the values involved are unsigned, so it's not a > signed/unsigned issue. We could give a better error message, but the > reality is if we can't trust the clock (and we can't really tell > between the clock skipping head briefly by a few weeks, and then set > back, and the case where the clock has been set back by two years), > e2fsck really should do a full check of the filesystem. > > - Ted -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
