Bug 1186066

Summary:	The docker stop operation doesn't work with --pid=host containers containing multiple processes
Product:	Red Hat Enterprise Linux 7	Reporter:	Steven Dake <sdake>
Component:	docker	Assignee:	Daniel Walsh <dwalsh>
Status:	CLOSED ERRATA	QA Contact:	atomic-bugs <atomic-bugs>
Severity:	high	Docs Contact:
Priority:	urgent
Version:	7.2	CC:	jkrieger, lsm5, lsu, steven.dake
Target Milestone:	rc	Keywords:	Extras
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2016-05-12 15:15:50 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	1186007

Description Steven Dake 2015-01-26 23:46:45 UTC

Description of problem:
docker stop blocks and fails to stop all processes in the container.

Version-Release number of selected component (if applicable):
1.4.1-beta2

How reproducible:
100%

Steps to Reproduce:
Make sure libvirtd, nova-compute, nova-network, and virtlockd is not running on the host.  If they are, disable them and reboot.

1. git clone https://github.com/sdake/compute-upgrade.git
2. cd compute-upgrade/tools
3. ./start-compute
4. Use docker ps to obtain docker container id for the nova-compute container
5. use docker stop <CONTAINER_ID> where container id is the container id from step #4

Actual results:
The operation blocks
killall -9 libvirtd
killall -9 virtlockd

This will unblock the docker stop operation

Expected results:
docker stop operation should terminate the libvirtd and virtlockd process and return within a few seconds.

Additional info:

Comment 1 Steven Dake 2015-01-26 23:48:57 UTC

Upstream issue tracker:

https://github.com/docker/docker/issues/10303

Comment 2 Steven Dake 2015-01-26 23:50:20 UTC

Cherry-pick of upstream patch that has been verified to resolve the problem:

https://github.com/rhatdan/docker/commit/0b4cade8b81cf8980977b5d353d7e6a0c08838b7

Comment 4 Daniel Walsh 2015-01-27 14:48:44 UTC

We have a build ready to go with this fix, but need the goahead on building a new package.

Comment 5 Daniel Walsh 2015-01-30 19:28:55 UTC

Fixed in 
docker-1.4.1-37.el7.x86_64.rpm

Comment 8 Steven Dake 2015-02-07 13:46:54 UTC

Daniel,

Any chance you could backport the latest rhel docker to fedora 21?  There is no --pid=host feature which would be nice for testing OpenStack upgrades.

Regards
-steve

Comment 9 Daniel Walsh 2015-02-09 10:25:18 UTC

That fix should be in docker-1.5 which I believe will be released tomorrow by docker. As soon as we get a released version we will update fedora 21 and Rawhide.  Will probably update f20 also.

Comment 11 Luwen Su 2016-05-01 08:10:59 UTC

This should work fine in docker-1.9.1-38.el7.x86_64, move to verified

# docker ps
CONTAINER ID        IMAGE                           COMMAND             CREATED             STATUS              PORTS               NAMES
de946711322b        sdake/fedora-rdo-nova-compute   "/start.sh"         32 seconds ago      Up 14 seconds                           cocky_lichterman
0f60869bd76e        sdake/fedora-rdo-nova-libvirt   "/start.sh"         45 seconds ago      Up 32 seconds                           focused_swirles
10d36abd5fa2        sdake/fedora-rdo-nova-network   "/start.sh"         49 seconds ago      Up 45 seconds                           berserk_wescoff
[root@dhcp-0-137 ~]# docker stop de946711322b
de946711322b

Comment 13 errata-xmlrpc 2016-05-12 15:15:50 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHSA-2016-1034.html