folders默认配置 shell_分布式存储Ceph RBD-Mirror灾备方案（二）镜像模式配置

接上文......

三、镜像模式配置

3.1 配置 rbd features 值- 对指定 rbd image 启用或禁用 features 属性

shell> rbd feature enable|disable {pool-name}/{image-name} {feature-name}

3.2 创建存储池

# 在主集群(以下简称 local )和备份集群(以下简称 remote )中分别执行

shell> ceph osd pool create 91donkey_image 128 128

pool '91donkey_image' created

shell> ceph osd pool application enable 91donkey_image rbd

enabled application 'rbd' on pool '91donkey_image'

3.3 启用存储池mirror功能

# 在 local 和 remote 两个集群中分别执行

# 语法：rbd mirror pool enable|disable {pool-name} {mode}

shell> rbd mirror pool enable 91donkey_image image

3.4 创建RBD

# local 集群

shell> rbd create -p 91donkey_image volume-312bf678-011a-11ea-a3ba-5254003111cb --size=10240

3.5 查看 local 集群的 image 是否具有 journaling 属性

开启 exclusive-lock 和 journaling 属性后才会记录 image 的事件，才可以被 rbd-mirror 检测到并同步到从集群

shell> rbd info 91donkey_image/volume-312bf678-011a-11ea-a3ba-5254003111cb

rbd image 'volume-312bf678-011a-11ea-a3ba-5254003111cb':

size 10 GiB in 2560 objects

order 22 (4 MiB objects)

id: 78a746b8b4567

block_name_prefix: rbd_data.78a746b8b4567

format: 2

features: layering, exclusive-lock, object-map, fast-diff, deep-flatten

op_features:

flags:

create_timestamp: Thu Nov 7 12:51:39 2019

3.6 禁用部分 features 属性

shell> rbd feature disable 91donkey_image/volume-312bf678-011a-11ea-a3ba-5254003111cb fast-diff

shell> rbd feature disable 91donkey_image/volume-312bf678-011a-11ea-a3ba-5254003111cb object-map

shell> rbd feature disable 91donkey_image/volume-312bf678-011a-11ea-a3ba-5254003111cb deep-flatten

# 再次查看 local 集群的 image 属性

shell> rbd info 91donkey_image/volume-312bf678-011a-11ea-a3ba-5254003111cb

rbd image 'volume-312bf678-011a-11ea-a3ba-5254003111cb':

size 10 GiB in 2560 objects

order 22 (4 MiB objects)

id: 78a746b8b4567

block_name_prefix: rbd_data.78a746b8b4567

format: 2

features: layering, exclusive-lock

op_features:

flags:

create_timestamp: Thu Nov 7 12:51:39 2019

# 开启 journaling 属性

shell> rbd feature enable 91donkey_image/volume-312bf678-011a-11ea-a3ba-5254003111cb journaling

3.7 增加同伴集群

把 local 集群和 remote 集群设为同伴，这个是为了让 rbd-mirror 进程找到它 peer 的集群存储池

# 语法： rbd mirror pool peer add {pool-name} {client-name}@{cluster-name}

# local 集群

shell> rbd mirror pool peer add 91donkey_image client.admin@remote

# remote 集群

shell> rbd mirror pool peer add 91donkey_image client.admin@local

备注：删除peer的语法，rbd mirror pool peer remove {pool-name} {peer-uuid}

3.8 查看 peer 的情况

- local 集群

- remote 集群

- 3.9 将镜像文件开启 mirroring 功能

shell> rbd mirror image enable -p 91donkey_image volume-312bf678-011a-11ea-a3ba-5254003111cb

- 3.10 查看 mirror pool 状态

shell> rbd mirror pool status 91donkey_image --verbose

- 3.11 集群间配置文件分发

# local 集群

shell> scp /etc/ceph/ceph.conf :/etc/ceph/local.conf

shell> scp /etc/ceph/ceph.client.admin.keyring :/etc/ceph/local.client.admin.keyring

# remote 集群

shell> scp /etc/ceph/ceph.conf :/etc/ceph/remote.conf

shell> scp /etc/ceph/ceph.client.admin.keyring :/etc/ceph/remote.client.admin.keyring

# local and remote 集群

shell> chown -R ceph.ceph /etc/ceph

- 3.12 开启 rbd-mirror 的同步进程

需要在两个集群上都配置，它会同时连接本地和远程的集群

shell> yum -y install rbd-mirror

# local 集群(-m 指定本集群的monitor ip，本例为10.20.58.191)

shell> rbd-mirror -m 10.20.58.191 --log-file=/var/log/ceph/ceph.rbd.mirror.log -i admin --setuser ceph --setgroup ceph

# remote 集群(-m 指定本集群的monitor ip，本例为10.20.50.71)

shell> rbd-mirror -m 10.20.50.71 --log-file=/var/log/ceph/ceph.rbd.mirror.log -i admin --setuser ceph --setgroup ceph

- 3.13 配置 rbd-mirror 开机启动

```

# 修改集群名字(如果集群自定义了名字，则需要修改)

# local 集群

shell> vim /usr/lib/systemd/system/ceph-rbd-mirror@.service

```默认集群名```

Environment=CLUSTER=ceph

```自定义的集群名```

Environment=CLUSTER=91donkey_local

shell> systemctl start ceph-rbd-mirror@admin

# remote 集群

shell> vim /usr/lib/systemd/system/ceph-rbd-mirror@.service

```默认集群名```

Environment=CLUSTER=ceph

```自定义的集群名```

Environment=CLUSTER=91donkey_remote

shell> systemctl start ceph-rbd-mirror@admin

主集群

mirror集群

作者：龍龍小宝