Bug#902117: [Debian-ha-maintainers] Bug#902117: corosync-qdevice will not daemonize/run
I'm seeing the same behavior in Ubuntu 18.04. What looks like early termination, and no debug output. The program is running though: $ sudo strace corosync-qdevice -df [...] openat(AT_FDCWD, "/dev/shm/qb-votequorum-event-12248-19618-30-header", O_RDWR) = 9 ftruncate(9, 8248) = 0 mmap(NULL, 8248, PROT_READ|PROT_WRITE, MAP_SHARED, 9, 0) = 0x7f22de722000 openat(AT_FDCWD, "/dev/shm/qb-votequorum-event-12248-19618-30-data", O_RDWR) = 10 ftruncate(10, 1052672) = 0 getpid()= 19618 sendto(11, "<30>Aug 9 11:06:16 corosync-qde"..., 102, MSG_NOSIGNAL, NULL, 0) = 102 mmap(NULL, 2105344, PROT_NONE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f22dac82000 mmap(0x7f22dac82000, 1052672, PROT_READ|PROT_WRITE, MAP_SHARED|MAP_FIXED, 10, 0) = 0x7f22dac82000 mmap(0x7f22dad83000, 1052672, PROT_READ|PROT_WRITE, MAP_SHARED|MAP_FIXED, 10, 0) = 0x7f22dad83000 close(10) = 0 close(9)= 0 sendto(8, "\20", 1, MSG_NOSIGNAL, NULL, 0) = 1 exit_group(1) = ? +++ exited with 1 +++
Bug#902117: [Debian-ha-maintainers] Bug#902117: corosync-qdevice will not daemonize/run
Control: severity -1 normal Control: tag -1 + upstream Valentin Vidic writes: > On Fri, Jul 06, 2018 at 12:50:42PM +0200, Ferenc Wágner wrote: > >> Thanks for the report. I've been pretty busy with other tasks, but I'll >> check this out as soon as possible, your report isn't forgotten. I ask >> for you patience till then. > > Feri, you still want to check this or should we close this issue? Hi Valentin, Sorry for the late response. I think that "corosync-qdevice -f -d" being completely silent is unexpected, thus a usability problem. There really should be a hint about this in the man page in my opinion. -- Regards, Feri
Bug#902117: [Debian-ha-maintainers] Bug#902117: corosync-qdevice will not daemonize/run
On Fri, Jul 06, 2018 at 12:50:42PM +0200, Ferenc Wágner wrote: > Thanks for the report. I've been pretty busy with other tasks, but I'll > check this out as soon as possible, your report isn't forgotten. I ask > for you patience till then. Feri, you still want to check this or should we close this issue? -- Valentin
Bug#902117: [Debian-ha-maintainers] Bug#902117: corosync-qdevice will not daemonize/run
On Fri, Jul 06, 2018 at 04:43:00PM -0400, Jason Gauthier wrote:
> Now, it's entirely possible that I do have a configuration issue
> causing corosync-qdevice to not start. However, the real issue is
> that corosync-qdevice does not log anything to stdout when run with
> "-f -d" (foreground, debug).
Just tried this on unstable and you are right there is no output
for "-f -d", but I do get this in the daemon.log:
Jul 7 14:29:51 sid1 corosync-qdevice[1507]: Configuring qdevice
Jul 7 14:29:51 sid1 corosync-qdevice[1507]: Can't read quorum.device.model
cmap key.
Jul 7 14:29:55 sid1 corosync-qdevice[1511]: Initializing votequorum
Jul 7 14:29:55 sid1 corosync-qdevice[1511]: shm size:1048589;
real_size:1052672; rb->word_size:263168
Jul 7 14:29:55 sid1 corosync-qdevice[1511]: shm size:1048589;
real_size:1052672; rb->word_size:263168
Jul 7 14:29:55 sid1 corosync-qdevice[1511]: shm size:1048589;
real_size:1052672; rb->word_size:263168
Jul 7 14:29:55 sid1 corosync-qdevice[1511]: Initializing local socket
Jul 7 14:29:55 sid1 corosync-qdevice[1511]: Registering qdevice models
Jul 7 14:29:55 sid1 corosync-qdevice[1511]: Configuring qdevice
Jul 7 14:29:55 sid1 corosync-qdevice[1511]: Can't read quorum.device.model
cmap key.
Maybe stdout does not exist for this service or you need to
tune this part of corosync.conf:
logging {
fileline: off
to_stderr: no
to_logfile: no
logfile:/var/log/corosync/corosync.log
to_syslog: yes
debug: off
timestamp: on
logger_subsys {
subsys: QUORUM
debug: off
}
}
--
Valentin
Bug#902117: [Debian-ha-maintainers] Bug#902117: corosync-qdevice will not daemonize/run
On Fri, Jul 6, 2018 at 10:20 AM Valentin Vidic wrote:
>
> On Fri, Jun 22, 2018 at 09:46:36AM -0400, Jason Gauthier wrote:
> > corosync-qdevice is a daemon that runs on each cluster node that help
> > provide a voting subsystem that utilizes corosync-qnet outside the
> > cluster.
> >
> > After installing the packages from debian stretch, and configuring the
> > application, it does not run. One can use -d and -f to troubleshoot
> > issues, and even in this situation no data is logged to the console,
> > or any syslog messages generated. The application immediately fails.
>
> corosync-qdevice is configured in corosync.conf, can you share the
> quorum block from there?
>
> If this is not configured I get the following error:
>
> Jul 6 16:08:08 node1 corosync-qdevice[2778]: Can't read quorum.device.model
> cmap key.
>
> But with a correct configuration it starts fine for me.
>
Interesting. Let me make some notes. First, on a 3rd node, I ran
corosync-qnetd, and if not setup then it gives you a nice bulleted
list to complete a certificate based set up. I did that, and
corosync-qnetd is launched and active on a 3rd node.
Now, it's entirely possible that I do have a configuration issue
causing corosync-qdevice to not start. However, the real issue is
that corosync-qdevice does not log anything to stdout when run with
"-f -d" (foreground, debug).
So, if there is a configuration problem I cannot troubleshoot it. If
I understand the issue correctly, this has something to do with libqb
and (perhaps) the way that it is built.
I grabbed the source code for corosync-qdevice and when I replaced the
logging functionality with a simple varargs based function I was able
to get some output and troubleshoot a little further.
Let me demonstrate:
root@alpha:/home/jgauthier# corosync-qdevice -f -d
root@alpha:/home/jgauthier#
This should result in some form of output. (This would be a great
test as for you as well)
Now, all that aside, my quorum block looks like this:
quorum {
# Enable and configure quorum subsystem (default: off)
# see also corosync.conf.5 and votequorum.5
# two_node: 1
provider: corosync_votequorum
expected_votes: 3
device {
votes: 1
model: net
net {
host: delta
}
}
delta is the host for corosync-qnetd.
Jason
Bug#902117: [Debian-ha-maintainers] Bug#902117: corosync-qdevice will not daemonize/run
On Fri, Jun 22, 2018 at 09:46:36AM -0400, Jason Gauthier wrote: > corosync-qdevice is a daemon that runs on each cluster node that help > provide a voting subsystem that utilizes corosync-qnet outside the > cluster. > > After installing the packages from debian stretch, and configuring the > application, it does not run. One can use -d and -f to troubleshoot > issues, and even in this situation no data is logged to the console, > or any syslog messages generated. The application immediately fails. corosync-qdevice is configured in corosync.conf, can you share the quorum block from there? If this is not configured I get the following error: Jul 6 16:08:08 node1 corosync-qdevice[2778]: Can't read quorum.device.model cmap key. But with a correct configuration it starts fine for me. -- Valentin
