Description of problem:
Current issue when customer is running a minor update on the compute nodes:
~~~ Hiding some information from the logs, since they contains customer private information, full error report will be added as a private comment[1]
[...]
"stderr: /usr/bin/docker-current: Error response from daemon: oci runtime error: container_linux.go:247: starting container process caused \"process_linux.go:245: running exec setns process for init caused \\\"exit status 17\\\"\".",
[...]
~~~

Previously, on the same update process, customer had issue on the controller nodes:
~~~ Hiding some information from the logs, since they contains customer private information, full error report will be added as a private comment[2]
[...]
        "stderr: /usr/bin/docker-current: Error response from daemon: service endpoint with name neutron_ovs_agent already exists."
~~~

He needed to re run the update 2 times, on each one a different controller showed the issue, finally the controllers were updated, but now the issue is on the compute, and that's a different one now.


I'm not sure both problems are related. In the current issue, do we have any workaround or solution? This issue is now blocking customer environment, he's not able to login to compute node nor create new instances.

Logs on the supportshell under /cases/02570698.

Version-Release number of selected component (if applicable):
OSP 13
ansible-tripleo-ipsec-8.1.1-0.20190513184007.7eb892c.el7ost
openstack-tripleo-common-8.7.1-2.el7ost
openstack-tripleo-common-containers-8.7.1-2.el7ost
openstack-tripleo-heat-templates-8.4.1-16.el7ost
openstack-tripleo-image-elements-8.0.3-1.el7ost
openstack-tripleo-puppet-elements-8.1.1-1.el7ost
openstack-tripleo-ui-8.3.2-3.el7ost
openstack-tripleo-validations-8.5.0-2.el7ost
puppet-tripleo-8.5.1-3.el7ost
python-tripleoclient-9.3.1-4.el7ost


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info: