Killing 1000 containers seems to be slow

== Comment: #0 - Thorsten Winkler - 2016-04-11 11:48:33 ==
Hi

to reproduce this I only started 1000 containers using:
     - docker run --name "sleep_$i" -d image_name sleep 100000

After that I tried to kill them in a for-loop with:
     - time docker kill sleep_$i

At the beginning of killing 1000 containers (see attachment) it takes on average about ~0.7/0.8 seconds to kill one container which seems for me a bit too slow. Maybe someone can have a look at this please?

#------------------------------------------------------------------------------------------------------------------------
Used system / system infos:
- 10GB Memory
- 4 CPUs

#docker version
Client:
 Version: 1.10.2
 API version: 1.22
 Go version: go1.6
 Git commit: c3959b1
 Built: Sun 13 Mar 2016 22:50:51 100
 OS/Arch: linux/s390x

Server:
 Version: 1.10.2
 API version: 1.22
 Go version: go1.6
 Git commit: c3959b1
 Built: Sun 13 Mar 2016 22:50:51 100
 OS/Arch: linux/s390x

#docker info
Containers: 0
 Running: 0
 Paused: 0
 Stopped: 0
Images: 1
Server Version: 1.10.2
Storage Driver: aufs
 Root Dir: /var/lib/docker/aufs
 Backing Filesystem: extfs
 Dirs: 2
 Dirperm1 Supported: true
Execution Driver: native-0.2
Logging Driver: json-file docker ve
Plugins:
 Volume: local
 Network: bridge null host
Kernel Version: 4.4.0-12-generic
Operating System: Ubuntu 16.04
OSType: linux
Architecture: s390x
CPUs: 16
Total Memory: 9.812 GiB
Name: s83lp36
ID: NV7U:5PTO:RDKL:JGJ2:OYXO:4VU6:THQQ:M5IS:FKDL:E5ZF:OBUZ:IELL

== Comment: #5 - Heinz-Werner Seeck - 2016-04-15 03:05:04 ==
The expected performance goal is around 0.4 seconds.