On Tue, 2006-03-21 at 02:40 +0100, Matthias Langer wrote: > On Mon, 2006-03-20 at 18:51 -0500, Willie Wong wrote: > > On Mon, Mar 20, 2006 at 04:51:31AM +0100, Matthias Langer wrote: > > > Recently i've got a power failure while one of my gentoo boxes was up > > > and running. Since then, the users command seems not to work correctly > > > anymore, as it claims for users to be logged in that almost certainly > > > aren't and don't have a single process running. Besides of this issue, > > > everthing seems to work as expected. Can anybody here tell me where to > > > look for the source of this problem ? > > > > > > Thanks, Matthias > > > > > > > I am running into a similar problem recently. I found out that sometimes, > > after > > updating 'system', init would restart. During an recent upgrade, after init > > restarts, the users that were logged-in at that time become "ghosts" of > > some sort. > > 'w' would show the correct number of people logged in, but some other > > commands > > won't. It might have something to do with the fact that wtmp is not > > registering the > > logouts from thost users. If I issue 'last | head' i would see something to > > that > > effect. > > > > I am wondering perhaps removing /var/log/wtmp would solve the issue (you > > might also > > want to touch /var/log/wtmp afterwards). It might require a rebooting > > (which I haven't > > gotten around to doing). > > Well, i tried that after booting into x86-2006.0-minimal. Unfortunatley, > that didn't solve my issuses; I found out that my system behaves after > the following pattern: > > login with user1: > $ users > user1 > $ exit > > login with user2: > $users > user1 user2 > $exit > > login with user3: > $users > user2 user3 > > Thus, the system seems to ignore logouts as long as there are not more > then two additional logins. > > I've also did a reiserfschk while staying in the live-cd environment, > but no corruptions where found ... > By the way, the output of $ w seems correct ...
-- gentoo-user@gentoo.org mailing list