mongodb数据库操作--备份 还原 导出 导入

• 首先数据库备份： mongodump -h IP --port 端口 -u 用户名 -p 密码 -d 数据库 -o 文件存在路径 

mongodump -h 127.0.0.1 -u admin -p xxx  -d blog -o '/home/timeless/桌面/mongodump' --authenticationDatabase admin

注意  --authenticationDatabase  参数制定认证数据库   否则会提示错误：

Failed: error connecting to db server: server returned error on SASL authentication step: Authentication failed.   //提示认证失败

也可以指定导出的集合  -collection

• 还原数据库实例

$ mongorestore -h XXX -u admin -p XXX -d blog  /home/timeless/桌面/mongodump/blog --authenticationDatabase admin
--25T15::40.481+    building a list of collections to restore from /home/timeless/桌面/mongodump/blog dir
--25T15::40.491+    reading metadata for blog.post from /home/timeless/桌面/mongodump/blog/post.metadata.json
--25T15::40.499+    reading metadata for blog.user from /home/timeless/桌面/mongodump/blog/user.metadata.json
--25T15::40.510+    reading metadata for blog.meta from /home/timeless/桌面/mongodump/blog/meta.metadata.json
--25T15::40.555+    restoring blog.meta from /home/timeless/桌面/mongodump/blog/meta.bson
--25T15::40.568+    restoring indexes for collection blog.meta from metadata
--25T15::40.602+    restoring blog.post from /home/timeless/桌面/mongodump/blog/post.bson
--25T15::40.635+    restoring blog.user from /home/timeless/桌面/mongodump/blog/user.bson
--25T15::40.637+    finished restoring blog.meta ( documents)
--25T15::40.637+    restoring indexes for collection blog.post from metadata
--25T15::40.645+    reading metadata for blog.page from /home/timeless/桌面/mongodump/blog/page.metadata.json
--25T15::40.646+    reading metadata for blog.content from /home/timeless/桌面/mongodump/blog/content.metadata.json
--25T15::40.646+    finished restoring blog.post ( document)
--25T15::40.646+    reading metadata for blog.option from /home/timeless/桌面/mongodump/blog/option.metadata.json
--25T15::40.646+    restoring indexes for collection blog.user from metadata
--25T15::40.679+    finished restoring blog.user ( documents)
--25T15::40.680+    reading metadata for blog.counters from /home/timeless/桌面/mongodump/blog/counters.metadata.json
--25T15::40.680+    restoring blog.page from /home/timeless/桌面/mongodump/blog/page.bson
--25T15::40.707+    restoring blog.option from /home/timeless/桌面/mongodump/blog/option.bson
--25T15::40.741+    restoring blog.content from /home/timeless/桌面/mongodump/blog/content.bson
--25T15::40.785+    restoring blog.counters from /home/timeless/桌面/mongodump/blog/counters.bson
--25T15::40.836+    restoring indexes for collection blog.option from metadata
--25T15::40.837+    restoring indexes for collection blog.page from metadata
--25T15::40.838+    restoring indexes for collection blog.content from metadata
--25T15::40.847+    finished restoring blog.page ( document)
--25T15::40.847+    reading metadata for blog.system.users from /home/timeless/桌面/mongodump/blog/system.users.metadata.json
--25T15::40.847+    finished restoring blog.content ( document)
--25T15::40.848+    reading metadata for blog.webinfo from /home/timeless/桌面/mongodump/blog/webinfo.metadata.json
--25T15::40.848+    restoring indexes for collection blog.counters from metadata
--25T15::40.848+    finished restoring blog.option ( document)
--25T15::40.884+    restoring blog.webinfo from /home/timeless/桌面/mongodump/blog/webinfo.bson
--25T15::40.884+    restoring indexes for collection blog.webinfo from metadata
--25T15::40.916+    finished restoring blog.counters ( documents)
--25T15::40.916+    restoring blog.system.users from /home/timeless/桌面/mongodump/blog/system.users.bson
--25T15::40.916+    finished restoring blog.webinfo ( documents)
--25T15::40.917+    restoring indexes for collection blog.system.users from metadata
--25T15::40.928+    finished restoring blog.system.users ( documents)
--25T15::40.928+    done

该也可以指定要恢复的集合  -collection

注意的地方 

• mongoimport 执行导入前，新数据库中没有 XX 表一样可以导入成功；
• mongoexport 和 mongoimport 所需的各个参数输入的先后顺序不严格；
• 对于副本集，mongoexport 的数据导出可以是副本集中的 primary 节点，也可以是 secondary 节点；
• 对于副本集，mongoimport 的数据导入必须是 primary 节点；
• 对于大数据量的导出，用 mongoexport 操作完全不会有生产问题；
• 对于大数据量的导入(超过 500 MB)，用 mongoimport 操作的话会很容易将 primary 节点拖垮，因为这会导致全量同步导致主节点崩溃，所以要避免生产环境中大数据量的导入。正确的做法是先建立单点，导入，之后再建立副本集。