LSF + Docker

docker 18.09.5

Kernel Version: 3.10.0-862.11.6.el7.x86_64

lsf 10.1.0.6

OS CentOs 7.6.1810

1.conf/lsf.conf

• 添加/修改

LSF_PROCESS_TRACKING=Y

LSF_LINUX_CGROUP_ACCT=Y

LSB_RESOURCE_ENFORCE="cpu memory"

2.conf/lsf.shared

• 添加

docker     Boolean ()       ()          (Docker container)

3.conf/lsf.cluster

• 添加

$your-host-name   !   !   1  3.5 () () (docker) 

4./conf/lsbatch/$clustername/configdir/lsb.applications

• 添加

Begin Application

NAME = app1

CONTAINER = docker[image(ubuntu:latest) options(--rm --network=host --ipc=host -v /etc/passwd:/etc/passwd  -v /etc/group:/etc/group) starter(root)]

DESCRIPTION = Test Docker Application Profile 1

End Application

5.badmin reconfig 验证是否可用

在非 root 用户下, bsub -app app1 -I cat /etc/lsb-release

DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=18.04
DISTRIB_CODENAME=bionic
DISTRIB_DESCRIPTION="Ubuntu 18.04.2 LTS"

1.badmin reconfig 出现 bad keyword CONTAINER:

• lsf 版本过低,升级 lsf 版本

2.修改之后未发现 docker 容器

• 首先确保 docker run 可用
• 确定 application 中的 EXEC_DRIVER 路径正确
• 确定 lsf.shared 中配置没有填错

https://www.ibm.com/support/knowledgecenter/SSWRJV_10.1.0/lsf_config_ref/lsb.applications.5.html?view=embed#lsb_applications_exec_driver

https://www.ibm.com/support/knowledgecenter/SSWRJV_10.1.0/lsf_docker/lsf_docker_config.html