[MongoDB Learning Notes 29] GridFS storage depth of MongoDB

The principle of GridFS is to divide a large file into several larger blocks and store each block as a separate document.

(1) blocks in GridFS are stored in a dedicated collection. Default is fs.chunks.

(2) in addition to storing each block of a file separately, you also need to put the meta-information of each block into a document store and set these documents in the fs.files collection by default.

For example, add an foo.txt document to the database foo:

[root@localhost] # echo "hello world" > foo.txt [root@localhost] # mongofiles-d foo put foo.txt connected to: 127.0.0.1 added file: {_ id: ObjectId ('54b3d62983047a88669bc529'), filename: "foo.txt", chunkSize: 261120, uploadDate: new Date (1421071914003), md5: "6f5902ac237024bdd0c176cb93063dc4", length: 12} done!

View the appropriate collection:

> show collections fs.chunks fs.files system.indexes >

View this document:

[root@localhost ~] # mongo foo MongoDB shell version: 2.6.6 connecting to: foo > db.fs.chunks.findOne () {"_ id": ObjectId ("54b51dc15caeb0e1a8722e8d"), "files_id": ObjectId ("54b51dc0c3ab6ae7c08f6028"), "n": 0, "data": BinData (0, "aGVsbG8gd29ybGQK")} >

Files_id: meta information of the file to which the block belongs

N: relative position of the block in the file

Data: binary data contained in a block

> db.fs.files.findOne () {"_ id": ObjectId ("54b51dc0c3ab6ae7c08f6028"), "filename": "foo.txt", "chunkSize": 261120, "uploadDate": ISODate ("2015-01-13T13:29:37.525Z"), "md5": "6f5902ac237024bdd0c176cb93063dc4", "length": 12} >

_ id: the unique ID of the file, which is consistent with the value of files_id in each block of the file

Length: the number of bytes contained in the file

ChunkSize: the size of each block that makes up the file, in bytes

UpdateDate: the time when the file was uploaded to GridFS

MD5: MD5 value of the file, calculated by the server