On 02/17/2012 09:08 AM, Patrick Horgan wrote:
On 02/17/2012 08:43 AM, Boris Epstein wrote:
On Fri, Feb 17, 2012 at 11:41 AM,
Patrick Horgan <phorgan1@xxxxxxxxx>
wrote:
# uname
-r
2.6.41.10-3.fc15.x86_64
First there was an upgrade:
Feb 12 05:09:27 s3 yum[14291]: Updated:
systemd-units-26-16.fc15.x86_64
Feb 12 05:09:43 s3 systemd[1]: Reexecuting.
Feb 12 05:09:44 s3 systemd[1]: systemd 26 running in system
mode. (+PAM
+LIBWRAP +AUDIT +SELINUX +SYSVINIT +LIBCRYPTSETUP; fedora)
Feb 12 05:09:44 s3 yum[14291]: Updated:
systemd-26-16.fc15.x86_64
Feb 12 05:09:45 s3 yum[14291]: Updated:
systemd-sysv-26-16.fc15.x86_64
Feb 12 05:09:50 s3 systemd[1]: Reloading.
Feb 12 05:09:32 s3 yum[14291]: Updated:
kernel-headers-2.6.42.3-2.fc15.x86_64
Feb 12 05:09:40 s3 yum[14291]: Installed:
kernel-2.6.42.3-2.fc15.x86_64
Then, when a cron job noticed the new kernel and tried to do
a shutdown
timed reboot via
/sbin/shutdown -r +5
in /var/log/messages I get:
Feb 13 03:00:01 s3 systemd-shutdownd[2733]: Received message
without
credentials. Ignoring.
and the system does not reboot. Since it talked about
credentials, I
double checked that selinux is not active on the machine:
# /usr/sbin/sestatus
SELinux status: disabled
yum update is done every day, but the problem still
persists. The cron
job notices every day that the running kernel is not the new
one and
tries to schedule a reboot and the same error message gets
into the logs.
I have been reading about systemd. I've read everything on
the wiki, the
man pages, and on Lennart's blog. I've watched videos of
his
presentation about systemd and I STILL don't have a clue.
My next
choice is to use strace on the pid and see what happens when
I try the
shutdown so I can see what it's trying to do that makes is
issue the log
message, but I was hoping someone here would know;)
Patrick
... elision of list sig by patrick ...
Interesting. Does
# init 5
or
# reboot
work?
Boris.
reboot works, and I'm assuming it's because it doesn't have to use
the delayed shutdown service.
... elision of message about not being able to try
reboot by patrick...
I was just reading the source for systemd, and the message is
issued here:
if (msghdr.msg_controllen < CMSG_LEN(sizeof(struct
ucred)) ||
control.cmsghdr.cmsg_level != SOL_SOCKET ||
control.cmsghdr.cmsg_type != SCM_CREDENTIALS ||
control.cmsghdr.cmsg_len != CMSG_LEN(sizeof(struct
ucred))) {
log_warning("Received message without credentials.
Ignoring.");
return 0;
}
They're getting credentials so that they can check to see if it's
uid 0 and if not fail. For some reason though, one of the things
that they're checking is not true. There's no way to tell which
from the message.
Still stuck. Another data point. Attaching to process 1 with
strace yields this:
open("/sys/fs/cgroup/systemd/system/systemd-shutdownd.service/tasks",
O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
open("/sys/fs/cgroup/systemd/system/systemd-shutdownd.service",
O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = -1 ENOENT (No such file
or directory)
sendmsg(14, {msg_name(0)=NULL,
msg_iov(2)=[{"l\4\1\1&\0\0\0\1\0\0\0o\0\0\0\1\1o\0\37\0\0\0/org/fre"...,
128}, {"!\0\0\0/system/systemd-shutdownd.se"..., 38}],
msg_controllen=0, msg_flags=0}, MSG_NOSIGNAL) = 166
We have another machine with the same hardware and software and it
weathered the same update fine and continues to automagically reboot
in the wee hours of the morning if it gets a new kernel.
Sigh.
Patrick
|
--
users mailing list
users@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
