我一直在尝试通过ansible部署ceph集群。当我试图将deploy_ceph_cluster.sh.j2呈现到shell脚本中时，我遇到了一些问题。为了更好地说明，我将提供一个最小的工作示例。
这是我的库存文件：
# inventory
[local]
localhost ansible_host=127.0.0.1 ansible_connection=local
[ceph_osd]
node1 ansible_host=192.168.1.2 rules="{'rule1': {'bcache': 'none', 'disks': '["/dev/sda", "/dev/sdc", "/dev/sdd", "/dev/sde"]'}}"
node2 ansible_host=192.168.1.3 rules="{'rule1': {'bcache': 'none', 'disks': '["/dev/sda", "/dev/sdc", "/dev/sdd", "/dev/sde"]'}}"
node3 ansible_host=192.168.1.4 rules="{'rule1': {'bcache': 'none', 'disks': '["/dev/sda", "/dev/sdc", "/dev/sdd", "/dev/sde"]'}}"
我的jinja2模板文件：
#!/bin/sh
createOSD() {
{% for host in groups['ceph_osd'] %}
{% for rule in hostvars[host].rules.values() %}
{% if rule.bcache != 'none' %}
ssh {{ host }} make-bcache -B {{ rule.disks | join(' ') }} -C {{ rule.bcache }} --wipe-bcache
bcache_name=$(lsblk -o KNAME {{ rule.bcache }} | grep bcache)
ssh {{ host }} echo writeback > /sys/block/$bcache_name/bcache/cache_mode
ceph-deploy osd create --data /dev/$bcache_name {{ host }}
{% else %}
echo {{rule.disks}}
{% for disk in rule.disks %}
ceph-deploy osd create --data {{ disk }} {{ host }}
{% endfor %}
{% endif %}
{% endfor %}
{% endfor %}
}
createOSD
最后，我的遗憾yml如下：
# deploy.yml file
---
- hosts: 127.0.0.1
connection: local
gather_facts: False
name: render jinja2 template
tasks:
- name: render all template_out template files in local
template: src=./deploy_ceph_cluster.sh.j2 dest=deploy_ceph_cluster.sh
...
您可以使用以下命令来呈现我的模板文件：
$ ansible-playbook deploy.yml -i inventory
PLAY [render jinja2 template] **************************************************
TASK [render all template_out template files in local] *************************
ok: [localhost]
PLAY RECAP *********************************************************************
localhost : ok=1 changed=0 unreachable=0 failed=0 skipped=0 rescued=0 ignored=0
到目前为止还不错。但是deploy_ceph_cluster.sh的上下文是错误的。其上下文如下：
#!/bin/sh
createOSD() {
echo [/dev/sda, /dev/sdc, /dev/sdd, /dev/sde]
ceph-deploy osd create --data [ node1
ceph-deploy osd create --data / node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data e node1
ceph-deploy osd create --data v node1
ceph-deploy osd create --data / node1
ceph-deploy osd create --data s node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data a node1
ceph-deploy osd create --data , node1
ceph-deploy osd create --data node1
ceph-deploy osd create --data / node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data e node1
ceph-deploy osd create --data v node1
ceph-deploy osd create --data / node1
ceph-deploy osd create --data s node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data c node1
ceph-deploy osd create --data , node1
ceph-deploy osd create --data node1
ceph-deploy osd create --data / node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data e node1
ceph-deploy osd create --data v node1
ceph-deploy osd create --data / node1
ceph-deploy osd create --data s node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data , node1
ceph-deploy osd create --data node1
ceph-deploy osd create --data / node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data e node1
ceph-deploy osd create --data v node1
ceph-deploy osd create --data / node1
ceph-deploy osd create --data s node1
ceph-deploy osd create --data d node1
ceph-deploy osd create --data e node1
ceph-deploy osd create --data ] node1
echo [/dev/sda, /dev/sdc, /dev/sdd, /dev/sde]
ceph-deploy osd create --data [ node2
ceph-deploy osd create --data / node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data e node2
ceph-deploy osd create --data v node2
ceph-deploy osd create --data / node2
ceph-deploy osd create --data s node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data a node2
ceph-deploy osd create --data , node2
ceph-deploy osd create --data node2
ceph-deploy osd create --data / node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data e node2
ceph-deploy osd create --data v node2
ceph-deploy osd create --data / node2
ceph-deploy osd create --data s node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data c node2
ceph-deploy osd create --data , node2
ceph-deploy osd create --data node2
ceph-deploy osd create --data / node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data e node2
ceph-deploy osd create --data v node2
ceph-deploy osd create --data / node2
ceph-deploy osd create --data s node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data , node2
ceph-deploy osd create --data node2
ceph-deploy osd create --data / node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data e node2
ceph-deploy osd create --data v node2
ceph-deploy osd create --data / node2
ceph-deploy osd create --data s node2
ceph-deploy osd create --data d node2
ceph-deploy osd create --data e node2
ceph-deploy osd create --data ] node2
echo [/dev/sda, /dev/sdc, /dev/sdd, /dev/sde]
ceph-deploy osd create --data [ node3
ceph-deploy osd create --data / node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data e node3
ceph-deploy osd create --data v node3
ceph-deploy osd create --data / node3
ceph-deploy osd create --data s node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data a node3
ceph-deploy osd create --data , node3
ceph-deploy osd create --data node3
ceph-deploy osd create --data / node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data e node3
ceph-deploy osd create --data v node3
ceph-deploy osd create --data / node3
ceph-deploy osd create --data s node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data c node3
ceph-deploy osd create --data , node3
ceph-deploy osd create --data node3
ceph-deploy osd create --data / node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data e node3
ceph-deploy osd create --data v node3
ceph-deploy osd create --data / node3
ceph-deploy osd create --data s node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data , node3
ceph-deploy osd create --data node3
ceph-deploy osd create --data / node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data e node3
ceph-deploy osd create --data v node3
ceph-deploy osd create --data / node3
ceph-deploy osd create --data s node3
ceph-deploy osd create --data d node3
ceph-deploy osd create --data e node3
ceph-deploy osd create --data ] node3
}
createOSD
正如您所看到的，每个磁盘变量都像[/dev/sda, /dev/sdc, /dev/sdd, /dev/sde]，但在下面的代码中，它被拆分为字符，例如，/dev/sda被拆分为"/"、"d"、"e"、"v"balabala。。。
createOSD() {
{% for host in groups['ceph_osd'] %}
{% for rule in hostvars[host].rules.values() %}
... omit some irrelevant code
{% for disk in rule.disks %}
ceph-deploy osd create --data {{ disk }} {{ host }}
{% endfor %}
{% endfor %}
{% endfor %}
}
我做错什么了吗？有人能给我一些提示或建议吗？提前谢谢。
{..., 'disks': '["/dev/sda", "/dev/sdc", "/dev/sdd", "/dev/sde"]'}
键'disks'的值在引号中，因此它实际上是str而不是list。您需要删除外部引号：
{..., 'disks': ['/dev/sda', '/dev/sdc', '/dev/sdd', '/dev/sde']}