Forced reboot just kills processes without shutting units/services down. In case of NetworkManager.service terminating of NM is racing with that of dhclient. If dhclient dies first NM brings down respective connection which - if used by NFS mounted in dracut (e.g. for repo=nfs: repo=nfsiso: installs) - causes hang when systemd umounts filesystems. Resolves: rhbz#818581 This patch obsoletes my first take on the bz: https://www.redhat.com/archives/anaconda-devel-list/2012-May/msg00271.html that didn't fix the issue completely. The hang kept occuring in some cases (repo=nfsiso: with kvm direct kernel boot on f17). --- anaconda | 8 ++++---- 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/anaconda b/anaconda index 939262d..8693588 100755 --- a/anaconda +++ b/anaconda @@ -71,13 +71,13 @@ def exitHandler(anaconda, exitCode=None): dracut_eject(drive.path) if anaconda.ksdata.reboot.action == KS_SHUTDOWN: - os.system("systemctl --force --no-wall poweroff") + os.system("systemctl --no-wall poweroff") elif anaconda.ksdata.reboot.action == KS_WAIT: - os.system("systemctl --force --no-wall halt") + os.system("systemctl --no-wall halt") else: # reboot action is KS_REBOOT or None - os.system("systemctl --force --no-wall reboot") + os.system("systemctl --no-wall reboot") elif not flags.imageInstall: - os.system("systemctl --force --no-wall reboot") + os.system("systemctl --no-wall reboot") def startMetacityWM(): childpid = os.fork() -- 1.7.4 _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list
