-
Notifications
You must be signed in to change notification settings - Fork 0
/
server.yml
34 lines (30 loc) · 1.13 KB
/
server.yml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
---
# This playbook deploys the whole application stack.
- name: deploy aliyun gpu tensorflow server
hosts: server
remote_user: root
vars:
hostname_domain: mytensor.com
hostname_name: dev.mytensor.com
timezone_zone: Asia/Shanghai
upload_cuda_cudnn_deb: false
cuda_version: 10.0.130-410.48
cuda_repo: cuda-repo-ubuntu1804-10-0-local-10.0.130-410.48_1.0-1_amd64
cudnn_repo: libcudnn7_7.4.2.24-1+cuda10.0_amd64
cudnn_dev_repo: libcudnn7-dev_7.4.2.24-1+cuda10.0_amd64
cuda_url: https://******.oss-cn-hangzhou-internal.aliyuncs.com/ml/cuda-repo-ubuntu1804-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
cudnn_url: https://******.oss-cn-hangzhou-internal.aliyuncs.com/ml/libcudnn7_7.4.2.24-1%2Bcuda10.0_amd64.deb
cudnn_dev_url: https://******.oss-cn-hangzhou-internal.aliyuncs.com/ml/libcudnn7-dev_7.4.2.24-1%2Bcuda10.0_amd64.deb
data_files:
- https://******.oss-cn-hangzhou-internal.aliyuncs.com/ml/abc.tar.gz
locale_enabled: yes
locale_locales:
- en_US.UTF-8
- zh_CN.UTF-8
roles:
- hostname
- timezone
- Stouts.locale
- common
- nvidia
- tensorflow