On Sun, Jul 05, 2015 at 12:35:57PM +0200, Giorgio Pioda wrote: > On Sun, Jul 05, 2015 at 12:28:00PM +0200, Andreas B. Mundt wrote: > > Hi, > > > > On Sun, Jul 05, 2015 at 11:40:36AM +0200, Giorgio Pioda wrote: > > > > > I can confirm a boot race condition (IIRC somebody talked six > > > months ago about autofs/systemd issues in this mailing list) > > > > > > Restarting manually the sercvices in the (more or less correct) > > > order on tjener: > > > > > > 1) nscd & nslcd > > > 2) nfs-common nfs-kernel > > > 3) autofs > > > > > > Fixes WS login > > > > > > I guess some careful upstream check is really needed > > > > Perhaps <URL:https://bugs.debian.org/759544>? > > > > It's probably still an issue in stable. > > Yes, this is it, I guess. > > Not 1 in 5 reboot, probably 4 on 5. > > Tomorrow I'll reboot and see if it get stuck again. In a VirtualBox test environment (combined server, gateway, workstation and in addition five diskless workstations which are forced to use KDM and NFS by setting DEFAULT_DISPLAY_MANAGER=/usr/bin/kdm in lts.conf) I've rebooted the main server 7 times; logging into randomly chosen workstations was successful each time. As reported, I've disabled nscd netgroup caching. This might serve as a workaround until a real fix for the problem is available. On tjener: service nscd stop rm /var/cache/nscd/netgroup sed -i '/netgroup/ s/yes/no/' /etc/nscd.conf service nscd start service nslcd restart service nfs-kernel-server restart Wolfgang
Attachment:
signature.asc
Description: Digital signature