Closed psfales closed 9 years ago
Hi,
I'm trying to reproduce your problem on centos 7 with docker 1.6 with some simple uses of weave, but I haven't succeeded so far. Could you provide the commands you use to get that "Error response from daemon"?
Thanks.
It's actually quite simple to reproduce. When I sent my earlier mail, I was running the powerstrip plugin which must have hidden the error.
I see noow that I can reproduce it with something as simple as
# weave launch
time="2015-05-14T17:55:41-04:00" level=fatal msg="Error response from daemon: : exit status 2"
I tried this on 4 different machines, all running CentOS7 and all recently upgraded to the recent docker-1.6.0-11.0.1.el7.centos.x86_64
I thought it might be an out-of-date weave, beacause "weave setup" pulled some new stuff, but it still fails the same way.
But I can do this!
# service docker stop
Redirecting to /bin/systemctl stop docker.service
# cp /usr/bin/docker-1.6-rc6 /usr/bin/docker
cp: overwrite `usr/bin/docker'? y
# service docker start
Redirecting to /bin/systemctl start docker.service
# weave launch
b06f12b1a9d0b7154e6f4b09ad11dfc4b4d5eb6ca4d3d87d5bed14152a22f356
#
Peter Fales Peter "at" fales-lorenz.net
On Thu, May 14, 2015 at 02:43:46PM -0700, David Wragg wrote:
Hi,
I'm trying to reproduce your problem on centos 7 with docker 1.6 with some simple uses of weave, but I haven't succeeded so far. Could you provide the commands you use to get that "Error response from daemon"?
Thanks.
-- Reply to this email directly or [1]view it on GitHub.
References
I'm still not able to reproduce that error: With weave 0.10.0 on centos7, weave launch
is successful for me. Which version of weave are you using?
Most likely the problem is related to selinux. I assume that /etc/selinux/config
shows that you have selinux set to enforcing and targeted? If you set selinux to permissive with setenforce 0
does the problem go away?
It might be worth doing fixfiles -R docker verify
to check that all the files in your docker install have the correct selinux attributes (though don't worry if it reports files under /var/lib/docker/containers/
).
I would guess that the RH-built docker is patched with enhanced selinux support, whereas your build was not, accounting for the difference in behaviour.
OK, Thanks and sorry for the trouble. Looks like the issue is my old version of weave.
I was under the mistaken impression that "weave setup" would upgrade me to the newest version of weave. I had tried running "weave setup" and it pulled new docker images, but I didn't get the new version of weave until I pulled the update from https://github.com/weaveworks/weave/releases/download/latest_release/weave
Now it's working for me with the CentOS7 version of docker-1.6!
(For the record, I have not been using selinux on any of our systems, so I don't think that's a factor)
Peter Fales Peter "at" fales-lorenz.net
I'm hoping you can help me debug why weave is not working with the stock version of docker-1.6
With the previous docker-1.5, there was a known issue with RHEL7/CentOS7 where docker would not properly mount a socket as a volume. (When we use "-v /var/run/docker.sock:/var/run/docker.sock") the file would not appear as a socket inside the container.
To work around that, I had installed a private copy of the docker-1.6 release candidate. I was hoping that the new 1.6 version would fix the problems. In fact, it does seem to fix that particular issue. But weave still does not work. I can start the weave container ok, but any "docker run" command reports: time="2015-05-14T17:06:32-04:00" level=fatal msg="Error response from daemon: : exit status 2"
If I go back to my private copy of the non-redhat docker-1.6-rc5 then it works!