Hi, On 1/17/13, m.roth@xxxxxxxxx <m.roth@xxxxxxxxx> wrote: > Hi, upen, > > upen wrote: >> >> On Thu, Jan 17, 2013 at 2:01 PM, <m.roth@xxxxxxxxx> wrote: >>> Odd problem that *just* surfaced. I updated a user to FC 17 - actually, >>> I >>> just built if for him new a couple of weeks ago. One issue that's >>> surfaced >>> is every so often, he comes to me to complain he can't log on. I restart >>> autofs, and everything's fine. >>> >>> All home directories are NFS mounted. >>> >>> systemctl is-enabled autofs.service >>> enabled >>> >>> I don't see anything that I can point to in the logs or dmesg. My user >>> goes away, leaves his workstation locked, comes back, and can't log in. >>> >>> selinux is permissive. >>> >>> Any thoughts on where I can start looking? >> >> Not sure how you leaned restarting autofs helps. > > I ssh'd into the box last time as me (I can go anywhere - I'm a sysadmin), > and saw that it couldn't find my home directory. I checked, and his wasn't > there, either. I tried restarting autofs, then su - back to myself, and I > was fine. Did it to him, ditto. Told him to try, and it worked. > <snip> >> I had seen similar problem when we were using nis/nis+ and user would > > Nope - we're using krb5 & AD. Nobody else is having this problem - AFAIK, > my manager isn't, nor is the one other FC box we have. Almost everything > else is CentOS. > >> lock their PC running gnome on RHEL 5 and when they came back and put >> their password the gnome screensaver would not authenticate >> successfully. I killed the gnome-screensaver process for those users >> after doing SSH to the box and the user would get to see their session >> immediately. May not be the same problem in your case but autofs >> wasn't the problem then. > > Hmmm... I'll look at that next time this happens. > <snip> >> One more thing you can do is enable debugging for autofs either in >> auto.master or by configuring sysconfig/autofs file > > One question on that: how noisy is that? All our linux boxes have > centralized logging, and if it spits out a ton of garbage, that's not > acceptable. If you can temporarily log locally then you can try that. Otherwise, try stopping autofs service and just manually mount the NFS share on the client using mount command. Check if this way things stay stable. There are plenty of autofs bugs out there you may have been hit by one those buggy versions out there :) https://bugzilla.redhat.com/buglist.cgi?query_format=specific&order=relevance+desc&bug_status=__open__&product=Fedora&content=autofs Thx. -- redhat-list mailing list unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list
