恢复configsvr副本集
准备目录
rm -rf /compile/cluster-restore/cfg*
mkdir -p /compile/cluster-restore/cfg1/data/db
mkdir -p /compile/cluster-restore/cfg1/log
mkdir -p /compile/cluster-restore/cfg2/data/db
mkdir -p /compile/cluster-restore/cfg2/log
mkdir -p /compile/cluster-restore/cfg3/data/db
mkdir -p /compile/cluster-restore/cfg3/log
操作步骤
- 准备单节点配置文件和数据目录,以单节点方式启动进程。
- 配置文件如下(restoreconfig/single_40303.yaml)。
net: bindIp: 127.0.0.1 port: 40303 unixDomainSocket: {enabled: false} processManagement: {fork: true, pidFilePath: /compile/cluster-restore/cfg1/configsvr.pid} storage: dbPath: /compile/cluster-restore/cfg1/data/db/ directoryPerDB: true engine: wiredTiger wiredTiger: collectionConfig: {blockCompressor: snappy} engineConfig: {directoryForIndexes: true, journalCompressor: snappy} indexConfig: {prefixCompression: true} systemLog: {destination: file, logAppend: true, logRotate: reopen, path: /compile/cluster-restore/cfg1/log/configsingle.log}
- 准备数据,将解压后的configsvr文件拷贝到单节点dbPath目录下。
/compile/download/backups/cac1efc8e65e42ecad8953352321bfeein02_41c8a32fb10245899708dea453a8c5c9no02/* /compile/cluster-restore/cfg1/data/db/
- 启动进程。
- 配置文件如下(restoreconfig/single_40303.yaml)。
- 连接单节点,执行配置命令。
连接命令:./mongo --host 127.0.0.1 --port 40303
- 执行如下命令,修改副本集配置信息。
var cf=db.getSiblingDB('local').system.replset.findOne();
cf['members'][0]['host']='127.0.0.1:40303';
cf['members'][1]['host']='127.0.0.1:40304';
cf['members'][2]['host']='127.0.0.1:40305';
cf['members'][0]['hidden']=false;
cf['members'][1]['hidden']=false;
cf['members'][2]['hidden']=false;
cf['members'][0]['priority']=1;
cf['members'][1]['priority']=1;
cf['members'][2]['priority']=1;
db.getSiblingDB('local').system.replset.remove({});
db.getSiblingDB('local').system.replset.insert(cf)
- 执行如下命令,清理内置账号。
db.getSiblingDB('admin').dropAllUsers();
db.getSiblingDB('admin').dropAllRoles();
- 执行如下命令,更新dds mongos和shard信息。
db.getSiblingDB('config').mongos.remove({});
先查询config.shards表中的多个shard的_id信息,用于下面语句的_id的查询条件。逐个更新每条记录。
db.getSiblingDB('config').shards.update({'_id' : 'shard_1'},{$set: {'host': 'shard_1/127.0.0.1:40306,127.0.0.1:40307,127.0.0.1:40308'}})
db.getSiblingDB('config').shards.update({'_id' : 'shard_2'},{$set: {'host': 'shard_2/127.0.0.1:40309,127.0.0.1:40310,127.0.0.1:40311'}})
db.getSiblingDB('config').mongos.find({});
db.getSiblingDB('config').shards.find({});
- 执行如下命令,关闭单节点进程。
- 执行如下命令,修改副本集配置信息。
- 搭建configsvr副本集。
- 准备副本集配置文件和目录,将configsvr1节点的dbPath文件拷贝到其他两个节点目录下。
cp -aR /compile/cluster-restore/cfg1/data/db/ /compile/cluster-restore/cfg2/data/db/
cp -aR /compile/cluster-restore/cfg1/data/db/ /compile/cluster-restore/cfg3/data/db/
- 修改configsvr-1节点配置文件,增加副本集配置属性(restoreconfig/configsvr_40303.yaml)。
net: bindIp: 127.0.0.1 port: 40303 unixDomainSocket: {enabled: false} processManagement: {fork: true, pidFilePath: /compile/cluster-restore/cfg1/configsvr.pid} replication: {replSetName: config} sharding: {archiveMovedChunks: false, clusterRole: configsvr} storage: dbPath: /compile/cluster-restore/cfg1/data/db/ directoryPerDB: true engine: wiredTiger wiredTiger: collectionConfig: {blockCompressor: snappy} engineConfig: {directoryForIndexes: true, journalCompressor: snappy} indexConfig: {prefixCompression: true} systemLog: {destination: file, logAppend: true, logRotate: reopen, path: /compile/cluster-restore/cfg1/log/configsvr.log}
- 启动进程。
- 修改configsvr-2节点配置文件,增加副本集配置属性(restoreconfig/configsvr_40304.yaml)。
net: bindIp: 127.0.0.1 port: 40304 unixDomainSocket: {enabled: false} processManagement: {fork: true, pidFilePath: /compile/cluster-restore/cfg2/configsvr.pid} replication: {replSetName: config} sharding: {archiveMovedChunks: false, clusterRole: configsvr} storage: dbPath: /compile/cluster-restore/cfg2/data/db/ directoryPerDB: true engine: wiredTiger wiredTiger: collectionConfig: {blockCompressor: snappy} engineConfig: {directoryForIndexes: true, journalCompressor: snappy} indexConfig: {prefixCompression: true} systemLog: {destination: file, logAppend: true, logRotate: reopen, path: /compile/cluster-restore/cfg2/log/configsvr.log}
- 启动进程
- 修改configsvr-3节点配置文件,增加副本集配置属性(restoreconfig/configsvr_40305.yaml)。
net: bindIp: 127.0.0.1 port: 40305 unixDomainSocket: {enabled: false} processManagement: {fork: true, pidFilePath: /compile/cluster-restore/cfg3/configsvr.pid} replication: {replSetName: config} sharding: {archiveMovedChunks: false, clusterRole: configsvr} storage: dbPath: /compile/cluster-restore/cfg3/data/db/ directoryPerDB: true engine: wiredTiger wiredTiger: collectionConfig: {blockCompressor: snappy} engineConfig: {directoryForIndexes: true, journalCompressor: snappy} indexConfig: {prefixCompression: true} systemLog: {destination: file, logAppend: true, logRotate: reopen, path: /compile/cluster-restore/cfg3/log/configsvr.log}
- 启动进程。
- 准备副本集配置文件和目录,将configsvr1节点的dbPath文件拷贝到其他两个节点目录下。
- 等待选主成功。
./mongo --host 127.0.0.1 --port 40303
执行命令rs.status(),查看是否已存在主节点Primary。