Plantilla PowerPoint

Download Report

Transcript Plantilla PowerPoint

MongoDB
Lugar y Fecha
Telefónica
I+D
Mongodb is...
A
document-oriented storage.
JSON-style
power.
documents
with
dynamic
schemas
offer
simplicity
Querying.
Rich, document-base queries.
Fast
In-Place Updates.
Atomic modifiers for contention-free performance.
Full
index Support.
Index on any attribute.
Auto-Sharding.
Scale horizontally without compromising funcionality.
Replication
& High Availability.
Mirror across LAN for scale.
Map/Reduce
Flexible (but limited) data processing.
Telefónica I+D
2
and
Mongodb is Not...
A
A
ACID database.
batch procesing platform.
Telefónica I+D
3
Mongodb Architecture
 MongoDB
is a document-oriented database, but this not implica
that it has the same concepts of RDMS, one of this concepts is
the “database”, within a MongoDB instance you can have zero or
more databases, each acting as high-level containers for
everything else
A
database can have zero or more “collections”, A collection
shares enough in common with a traditional table, a collection is
the same as a table with a indeterminate number of columns.
 Collections
is made up of zero or more documents.
document can safely be thought of as a “row”.
Again,
a
A
document is made up one or more fields, which yo can probable
guess are a lot like “columns”.
 Indexes
 Cursors
in MongoDB function much like their RDBMS counterparts.
are different than the other five concepts but they are
important enough. The import thing to understand about cursors is
that when you ask MongoDB for data, it returns a cursor, which we
can do things to, such as counting or skipping ahead, without
actually pulling down data.
Telefónica I+D
4
Mongodb Architecture
 Mongo
uses memory mapped files to access data, which results in
large numbers being displayed in tools like top for the mongod
process.
 Mongo
eliminates the need (in some cases) for a separate object
caching layer. Queries that result in file system RAM cache hits
are very fast as the object's representation in the database is
very close to its representation in application memory. Also, the
MongoDB can scale to any level and provides an object cache and
database integrated together, which is very helpful as there is
no risk of retrieving stale data from the cache. In addition, the
complex queries a full DBMS provides are also possible.
 MongoDB
supports
write-ahead
journaling
of
operations
facilitate fast crash recovery and durability in the storage
Telefónica I+D
5
to
Mongodb Api
Telefónica I+D
Data Type
Type Number
Double
1
String
2
Object
3
Array
4
Binary data
5
Object id
7
Boolean
8
Date
9
Null
10
Regular Expresion
11
6
Mongodb Api
 The
conection
to
the
databases
front
the
aplications are a propiatery protocol that runs over
TCP/IP in binary format. The apis have a control the
write behavior for with various options, as well as
exception raising on error conditions.
 NONE: No exceptions are
 NORMAL: Exceptions are
raised, even for network issues
raised for network issues, but
not server errors
 SAFE:
Exceptions are raised for network issues, and
server errors; waits on a server for the write
operation
 FSYNC_SAFE:
Exceptions are raised for network issues,
and server errors and the write operation waits for the
server to flush the data to disk
 REPLICAS_SAFE:
Exceptions
are raised for network
issues, and server errors; waits for at least 2 servers
for the write operation
Telefónica I+D
7
Mongodb Api
 Insert:
it is the basic method for adding data into
MongoDB. When we insert a new document into a
collection this will add an “_id” key to the
document before save it. When insert a new object
into the collection, the driver coverts the data
structure into BSON, which
it then sends to the
database. Then it saves the document, and regenerate
the _id index, and all the secondary index that the
new object has.
 db.collection.insert(doc);
 db.collection.save(doc); // updates if exists; inserts if new (_id)
 Delete: it is the basic method for deleting data.
 db.videos.remove( { rating : { $lt : 3.0 }, $atomic : true } )
Telefónica I+D
8
Mongodb Api
 Query:
MongoDB's support for dynamic (ad hoc)
queries. MongoDB supports a number of query objects
for fetching data.
 Db.collection.find(query,fields)
 Method:
 count()
 sort()
 limit()
 skip()
Telefónica I+D
9
Mongodb Api
Query Modifier Operations
$gt (>) , $gte(>=), $lt (<), $gle(<=)
db.collection.find({ "field" : { $gt: value } } );
$all
db.things.find( { a: { $all: [ 2, 3 ] } } );
$exists
db.things.find( { a : { $exists : true } } ); present
$mod
db.things.find( { a : { $mod : [ 10 , 1 ] } } )
$ne
db.things.find( { x : { $ne : 3 } } );
$in
db.things.find({j:{$in: [2,4,6]}});
$nin
db.things.find({j:{$nin: [2,4,6]}});
$or
db.foo.find( { $or : [ { a : 1 } , { b : 2 } ] } )
$size
db.things.find( { a : { $size: 1 } } );
$type
db.things.find( { a : { $type : 2 } } );
$elemMatch
t.find( { x : { $elemMatch : { a : 1}})
$regex
db.customers.find( { name : { $regex : 'acme.*corp', $options: 'i' } } );
$where
db.myCollection.find( { $where: "this.a > 3" } );
Telefónica I+D
10
Mongodb Api
 Update:
replaces the document matching criteria
entirely with objNew. If you only want to modify
some fields, you should use the atomic modifiers
below.
 db.collection.update( criteria, objNew, upsert, multi ).
 Arguments
 criteria - query which selects the record to update;
 objNew - updated object or $ operators (e.g., $inc) which manipulate
the object

upsert - if this should be an "upsert"; that is, if the record does
not exist, insert it

multi - if all documents matching criteria should be updated
Telefónica I+D
11
Mongodb Api
Update Modifier Operations
$inc
{ $inc : { field : value } }
$set
{ $set : { field : value } }
$unset
{ $unset : { field : 1} }
$push
{ $push : { field : value } }
$pushAll
{ $pushAll : { field : value_array } }
$addToSet
{ $addToSet : { field : value } }
$pop
{ $pop : { field : 1 } }
{ $pop : { field : -1 } }
$pull
{ $pull : { field : _value } }
$pullAll
{ $pullAll : { field : value_array } }
$rename
{ $rename : { old_field_name : new_field_name } }
$bit
{$bit : {field : {and : 5, or : 2}}}
$ ( positional number)
Telefónica I+D
{$inc:{'comments.$.votes':1}}
12
Mongodb Api
 Upserts
: it is a special type of update, in where
if no document is found to upate, a new document
will be created in a atomic operation.

db.jobs.findAndModify({query: {},update : {}, fields: {},
Arguments
Query
{}
Sort
{}
Remove
N/A
Update
N/A
New
False
Fields
All fields
upsert
False
Telefónica I+D
13
new: false } )
Mongodb Api
Example in console.
Telefónica I+D
14
Mongodb Index
These indexes are implemented as "B-Tree" indexes.
the indexes in MongoDB work a lot like indexes in
a
relational database: they help improve query and
sorting performance, the indexes are created manual,
and could be unique, ascending or descending.
Index on _id is always created and can`t be deleted.
I can index on a key inside of an embedded document.
Telefónica I+D
15
Mongodb Index
Example in console.
Telefónica I+D
16
Mongodb Sharding
MongoDB supports auto-sharding, and auto.balancing.
Sharding is an approach to scalability which
separates your data across multiple servers.
With sharding we instead scale horizontally to
achieve
the
same
computacional/storage/memory
footprint from smaller servers.
We will show the vertically scale db and the
horizontally scaled db for comparison.
A sharded collection has a shard key. The collection
is partitioned using this key. In this example we
uses ident.
 {ident:…, data:…}
Telefónica I+D
17
Mongodb Sharding
There are three components in sharding.
 Mongos: routing procesing.
 Config servers: store the metadata.
 Mongod: store the data.
Range
Shard
Ident in [-inf,200)
2
Ident in (200, 400]
4
…
Ident in [1000,inf)
Telefónica I+D
18
10
Mongodb Sharding
Metadata
is
maintained
on
chunks
represented by shard key ranges.
Range
which
Shard
Ident in [-inf,200)
2
Ident in (200, 400]
4
…
Ident in [1000,inf)
Telefónica I+D
19
are
10
Mongodb Sharding
Querys
with the shard key are sent to the node who
has this shard.
Query
Find({ident :{$gt 100, $lt:150})
Telefónica I+D
20
Shards
2
Mongodb Sharding
Querys
with the shard key and the data are in some
shard, it as sent to these nodes.
Query
Find({ident :{$gt 100, $lt:250})
Telefónica I+D
21
Shards
2,3
Mongodb Sharding
Querys
with the no shard key are sent to all nodes.
Query
Find({data:{$gt 100, $lt:250})
Telefónica I+D
22
Shards
2,3
Mongodb Sharding
Example in console.
Telefónica I+D
23
Mongodb Replication
MongoDB
replication
works
similarly
to
how
relational database replication works. Writes are
sent to a single server, the master, which then
synchronizes itself to one or more other servers,
the slaves. If the master goes down, a slave can be
promoted to act as the new master.
While replication can improve performance (by
distributing reads), its main purpose is to increase
reliability.
There are two types of replication:
 Master slaver replication.
 Replica sets.
Telefónica I+D
24
Mongodb Replication
Master
Setup
slaver replication.
master:
 bin/mongod –master
Setup slaver:
 bin/mongod --slave
Telefónica I+D
–oplogSize 100
--source <IP>[:<port>] --slavedelay 1 --autoresync
25
Mongodb Replication
Replica
sets.
 mongod --replSet “setname”
Shell in one node
cfg = {
... _id : “setname",
... members : [
... { _id : 0, host : “host1" },
... { _id : 1, host : “host2" },
... { _id : 2, host : “host3" } ] }
rs.initiate(cfg)
rs.status()
Options
arbiterOnly
buildIndexes
hidden
priority
slaveDelay
votes
tags
Telefónica I+D
26
Mongodb Map&Reduce
DB
is useful for batch
aggregation operations.
Telefónica I+D
27
processing
of
data
and
Mongodb Tools
Mongodump: export to binary data
Mongoexport: export to csv
Mongofiles: load data into mongodb
MongoImport: import txt, csv or json data
mongoRestore: mongorestore takes the output
from
mongodump: and restores it.
mongoSniff: This utility is to MongoDB what tcpdump
is to TCP/IP
Mongostat: Use the mongostat utility to quickly view
statistics on a running mongod instance
Telefónica I+D
28