mongodump.exe備份的原理是通過一次查詢獲取當前服務器快照，并將快照寫入磁盤中，因此這種方式保存的也不是實時的，因為在獲取快照后，服務器還會有數(shù)據(jù)寫入，為了保證備份的安全，同樣我們還是可以利用fsync鎖使服務器數(shù)據(jù)暫時寫入緩存中。

備份前的檢查
> show dbs
MyDB 0.0625GB
admin (empty)
bruce 0.0625GB
local (empty)
test 0.0625GB
> use MyDB
switched to db MyDB
> db.users.find()
{ "_id" : ObjectId("4e290aa39a1945747b28f1ee"), "a" : 1, "b" : 1 }
{ "_id" : ObjectId("4e2cd2182a65c81f21566318"), "a" : 3, "b" : 5 }
>

整庫備份：
mongodump -h dbhost -d dbname -o dbdirectory
-h：MongDB所在服務器地址，例如：127.0.0.1，當然也可以指定端口號：127.0.0.1:27017
-d：需要備份的數(shù)據(jù)庫實例，例如：test
-o：備份的數(shù)據(jù)存放位置，例如：c:\data\dump，當然該目錄需要提前建立，在備份完成后，系統(tǒng)自動在dump目錄下建立一個test目錄，這個目錄里面存放該數(shù)據(jù)庫實例的備份數(shù)據(jù)。

mongodump的官方說明(可通過mongodump --help查看)：
options:
--help          produce help message
-v [ --verbose ]     be more verbose (include multiple times for more
              verbosity e.g. -vvvvv)
--version        print the program's version and exit
-h [ --host ] arg    mongo host to connect to ( /s1,s2 for
              sets)
--port arg        server port. Can also use --host hostname:port
--ipv6          enable IPv6 support (disabled by default)
-u [ --username ] arg  username
-p [ --password ] arg  password
--dbpath arg       directly access mongod database files in the given
              path, instead of connecting to a mongod server -
              needs to lock the data directory, so cannot be used
              if a mongod is currently accessing the same path
--directoryperdb     if dbpath specified, each db is in a separate
              directory
--journal        enable journaling
-d [ --db ] arg     database to use
-c [ --collection ] arg collection to use (some commands)
-o [ --out ] arg (=dump) output directory or "-" for stdout
-q [ --query ] arg    json query
--oplog         Use oplog for point-in-time snapshotting
--repair         try to recover a crashed database
--forceTableScan     force a table scan (do not use $snapshot)

整庫恢復：
mongorestore -h dbhost -d dbname –directoryperdb dbdirectory
-h：MongoDB所在服務器地址
-d：需要恢復的數(shù)據(jù)庫實例，例如：test，當然這個名稱也可以和備份時候的不一樣，比如test2
–directoryperdb：備份數(shù)據(jù)所在位置，例如：c:\data\dump\test，這里為什么要多加一個test，而不是備份時候的dump，讀者自己查看提示吧！
–drop：恢復的時候，先刪除當前數(shù)據(jù)，然后恢復備份的數(shù)據(jù)。就是說，恢復后，備份后添加修改的數(shù)據(jù)都會被刪除，慎用哦！

mongorestore的官方說明(可通過mongorestore --help查看)：
options:
--help         produce help message
-v [ --verbose ]    be more verbose (include multiple times for more
             verbosity e.g. -vvvvv)
--version        print the program's version and exit
-h [ --host ] arg    mongo host to connect to ( /s1,s2 for sets)
--port arg       server port. Can also use --host hostname:port
--ipv6         enable IPv6 support (disabled by default)
-u [ --username ] arg  username
-p [ --password ] arg  password
--dbpath arg      directly access mongod database files in the given
             path, instead of connecting to a mongod server -
             needs to lock the data directory, so cannot be used
             if a mongod is currently accessing the same path
--directoryperdb    if dbpath specified, each db is in a separate
             directory
--journal        enable journaling
-d [ --db ] arg     database to use
-c [ --collection ] arg collection to use (some commands)
--objcheck       validate object before inserting
--filter arg      filter to apply before inserting
--drop         drop each collection before import
--oplogReplay      replay oplog for point-in-time restore
--oplogLimit arg    exclude oplog entries newer than provided timestamp
             (epoch[:ordinal])
--keepIndexVersion   don't upgrade indexes to newest version
--noOptionsRestore   don't restore collection options
--noIndexRestore    don't restore indexes
--w arg (=1)      minimum number of replicas per write

單個collection備份：
mongoexport -h dbhost -d dbname -c collectionname -f collectionKey -o dbdirectory
-h: MongoDB所在服務器地址
-d: 需要恢復的數(shù)據(jù)庫實例
-c: 需要恢復的集合
-f: 需要導出的字段(省略為所有字段)
-o: 表示導出的文件名

mongoexport的官方說明(可通過mongoexport --help查看)：
--help          produce help message
-v [ --verbose ]     be more verbose (include multiple times for more
              verbosity e.g. -vvvvv)
--version         print the program's version and exit
-h [ --host ] arg     mongo host to connect to ( /s1,s2 for
              sets)
--port arg        server port. Can also use --host hostname:port
--ipv6          enable IPv6 support (disabled by default)
-u [ --username ] arg   username
-p [ --password ] arg   password
--dbpath arg       directly access mongod database files in the given
              path, instead of connecting to a mongod server -
              needs to lock the data directory, so cannot be used
              if a mongod is currently accessing the same path
--directoryperdb     if dbpath specified, each db is in a separate
              directory
--journal         enable journaling
-d [ --db ] arg      database to use
-c [ --collection ] arg  collection to use (some commands)
-f [ --fields ] arg    comma separated list of field names e.g. -f
              name,age
--fieldFile arg      file with fields names - 1 per line
-q [ --query ] arg    query filter, as a JSON string
--csv           export to csv instead of json
-o [ --out ] arg     output file; if not specified, stdout is used
--jsonArray        output to a json array rather than one object per
              line
-k [ --slaveOk ] arg (=1) use secondaries for export if available, default
              true
--forceTableScan     force a table scan (do not use $snapshot)

單個collection恢復：
mongoimport -d dbhost -c collectionname –type csv –headerline –file
-type: 指明要導入的文件格式
-headerline: 批明不導入第一行，因為第一行是列名
-file: 指明要導入的文件路徑

mongoimport的官方說明(可通過mongoimport --help查看)：
--help         produce help message
-v [ --verbose ]    be more verbose (include multiple times for more
             verbosity e.g. -vvvvv)
--version        print the program's version and exit
-h [ --host ] arg    mongo host to connect to ( /s1,s2 for sets)
--port arg       server port. Can also use --host hostname:port
--ipv6         enable IPv6 support (disabled by default)
-u [ --username ] arg  username
-p [ --password ] arg  password
--dbpath arg      directly access mongod database files in the given
             path, instead of connecting to a mongod server -
             needs to lock the data directory, so cannot be used
             if a mongod is currently accessing the same path
--directoryperdb    if dbpath specified, each db is in a separate
             directory
--journal        enable journaling
-d [ --db ] arg     database to use
-c [ --collection ] arg collection to use (some commands)
-f [ --fields ] arg   comma separated list of field names e.g. -f name,age
--fieldFile arg     file with fields names - 1 per line
--ignoreBlanks     if given, empty fields in csv and tsv will be ignored
--type arg       type of file to import. default: json (json,csv,tsv)
--file arg       file to import from; if not specified stdin is used
--drop         drop collection first
--headerline      CSV,TSV only - use first line as headers
--upsert        insert or update objects that already exist
--upsertFields arg   comma-separated fields for the query part of the
             upsert. You should make sure this is indexed
--stopOnError      stop importing at first error rather than continuing
--jsonArray       load a json array, not one item per line. Currently
             limited to 16MB.

其他導入與導出操作：

1. mongoimport -d my_mongodb -c user user.dat

參數(shù)說明:

-d 指明使用的庫, 本例中為” my_mongodb”

-c 指明要導出的表, 本例中為”user”

可以看到導入數(shù)據(jù)的時候會隱式創(chuàng)建表結(jié)構(gòu)

2. mongoexport -d my_mongodb -c user -o user.dat

參數(shù)說明:

-d 指明使用的庫, 本例中為” my_mongodb”

-c 指明要導出的表, 本例中為”user”

-o 指明要導出的文件名, 本例中為”user.dat”

從上面可以看到導出的方式使用的是JSON 的樣式.