On 2009-09-29 10:58 +0200, Sven Joachim wrote: > Package: initscripts > Version: 2.87dsf-6 > Severity: normal > > Since I upgraded udev to version 146-3 on Sunday evening, the shutdown > process hiccups. One symptom is that wpa_supplicant does not get > killed, leaving /var and /usr busy; see #548694 for that issue. > > The other problem is that sendsigs takes ten seconds and then reports > failure. I tried running all the scripts in /etc/rc6.d manually and put > a "set -x" into the sendsigs script; the output is attached, as is the > output of "pstree" before and after running sendsigs.
Putting "ps -ef" calls into the sendsigs script, it seems that a dhclient3 process is the culprit. Inside the "for seq in 1 2 3 4 5 6 7 8 9 10" loop in the sendsigs script the process table looks like this:
UID PID PPID C STIME TTY TIME CMD root 1 0 0 16:08 ? 00:00:00 init [6] root 2 0 0 16:08 ? 00:00:00 [kthreadd] root 3 2 0 16:08 ? 00:00:00 [migration/0] root 4 2 0 16:08 ? 00:00:00 [ksoftirqd/0] root 5 2 0 16:08 ? 00:00:00 [migration/1] root 6 2 0 16:08 ? 00:00:00 [ksoftirqd/1] root 7 2 0 16:08 ? 00:00:00 [events/0] root 8 2 0 16:08 ? 00:00:00 [events/1] root 9 2 0 16:08 ? 00:00:00 [khelper] root 10 2 0 16:08 ? 00:00:00 [async/mgr] root 11 2 0 16:08 ? 00:00:00 [kblockd/0] root 12 2 0 16:08 ? 00:00:00 [kblockd/1] root 13 2 0 16:08 ? 00:00:00 [kacpid] root 14 2 0 16:08 ? 00:00:00 [kacpi_notify] root 15 2 0 16:08 ? 00:00:00 [ata/0] root 16 2 0 16:08 ? 00:00:00 [ata/1] root 17 2 0 16:08 ? 00:00:00 [ata_aux] root 18 2 0 16:08 ? 00:00:00 [kseriod] root 19 2 0 16:08 ? 00:00:00 [pdflush] root 20 2 0 16:08 ? 00:00:00 [pdflush] root 21 2 0 16:08 ? 00:00:00 [kswapd0] root 22 2 0 16:08 ? 00:00:00 [aio/0] root 23 2 0 16:08 ? 00:00:00 [aio/1] root 24 2 0 16:08 ? 00:00:00 [scsi_eh_0] root 25 2 0 16:08 ? 00:00:00 [scsi_eh_1] root 28 2 0 16:08 ? 00:00:00 [kpsmoused] root 579 2 0 16:08 ? 00:00:00 [ksuspend_usbd] root 613 2 0 16:08 ? 00:00:00 [khubd] root 698 2 0 16:08 ? 00:00:00 [hd-audio0] root 890 2 0 16:08 ? 00:00:00 [phy0] root 958 2 0 16:08 ? 00:00:00 [kjournald2] root 959 2 0 16:08 ? 00:00:00 [kjournald2] root 960 2 0 16:08 ? 00:00:00 [kjournald2] root 961 2 0 16:08 ? 00:00:00 [kjournald2] root 962 2 0 16:08 ? 00:00:00 [kjournald2] daemon 1098 1 0 16:08 ? 00:00:00 /sbin/portmap root 1229 1 0 16:08 ? 00:00:00 /usr/sbin/rsyslogd -c4 root 1512 2 0 16:08 ? 00:00:00 [rpciod/0] root 1513 2 0 16:08 ? 00:00:00 [rpciod/1] root 1546 1 0 16:08 ? 00:00:00 /sbin/wpa_supplicant -s -B -P /var/run/wpa_supplicant.wlan0.pid -i wlan0 -D wext -C /var/run/wpa_supplicant root 1682 1 0 16:08 ? 00:00:00 dhclient3 -pf /var/run/dhclient.wlan0.pid -lf /var/lib/dhcp3/dhclient.wlan0.leases wlan0 root 1828 1 0 16:30 ? 00:00:00 /bin/sh /etc/init.d/rc 6 root 2016 1828 0 16:30 ? 00:00:00 /bin/sh /etc/rc6.d/K03sendsigs stop root 2052 2016 0 16:31 ? 00:00:00 ps -ef
After the "killall5 -9 $OMITPIDS" call the dhclient3 process is still not quite dead and listed like this: root 1682 1 0 16:08 ? 00:00:00 [dhclient3] Any clue why killall5 did not manage to kill it for good? Sven
_______________________________________________ Pkg-sysvinit-devel mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/pkg-sysvinit-devel
