Shulou(Shulou.com)05/31 Report--

How to carry out the practice of SequoiaDB + JanusGraph, in view of this problem, this article introduces the corresponding analysis and solution in detail, hoping to help more partners who want to solve this problem to find a more simple and feasible method.

JanusGraph

In fact, in the field of graph data, Neo4j is really dominant, but its community version, performance "yan" (ge) has gone too far, and various functions have been cut, making it difficult to use in a production environment. As for the enterprise version, there is no specific budget for this area at present.

And look at the second in the picture database-JanusGraph, the top project of the Apache Foundation, continues to move forward with the halo of the star project Titan. "this kid must be something," that's what I think.

If you are curious about the past life of JanusGraph, you can find out what DataStax (the parent company of Cassandra) did to Titan. Then a group of awesome programmers who love open source work on their own. Anyway, this story sounds similar to the story of MySQL and MariaDB falling in love and killing each other, but the end of JanusGraph is even more heroic.

I found an overall architecture diagram from the official website of JanusGraph. You can see that the modules of JanusGraph are quite rich and the functions are relatively comprehensive.

Because of this separate architecture of computing and storage, application developers do not need to pay attention to the underlying data distributed implementation at all, as long as they concentrate on their own application logic.

Of course, if you are a DBA, you still need to learn how to configure the underlying distributed storage of SequoiaDB.

As mentioned earlier, because the design and interface of JanusGraph to the underlying storage is based on column storage, some adjustments need to be made when designing for SequoiaDB for JanusGraph.

The first is the family design in the cluster. I took it apart and mapped the different family directly to a Collection of SequoiaDB.

At first, I wanted to save all the columns of a record to a BSON in SequoiaDB, but later, because JanusGraph will rely on the column sorting function of HBase to return records, it is impossible to sort different columns of a record in SequoiaDB. So in the end, I split a record in JanusGraph into multiple BSON records in the following form. Records use RowKey to maintain the integrity of one of their records.

{RowKey: ", Key:", Value: ""} {RowKey: ", Key:", Value: "} {RowKey:", Key: ", Value:"} {RowKey: ", Key:", Value: ""}

In BSON, the data types of the RowKey, Key, and Value fields are all in Binary format, which is JanusGraph's own unique parsing method. The record saved by JanusGraph in SequoiaDB is shown in the following example:

{"_ id": {"$oid": "5e410c444f025855e5552b4c"}, "Key": {"$binary": "/ + x38ABZ40DXrgsGMwYTgxZmZiMTc2ODYtY2hlbjE =", "$type": "0"}, "RowKey": {"$binary": "AAAAAAAAAAM=", "$type": "0"}, "Value": {"$binary": "," $type ":" 0 "}}

02

JanusGraph code modification

Earlier, I showed you how JanusGraph recognizes different storage products, so to add the storage option of SequoiaDB database, you first need to add the SequoiaDB option to the StandardStoreManager class. The modified parts are as follows: BDB_JE ("org.janusgraph.diskstorage.berkeleyje.BerkeleyJEStoreManager", "berkeleyje"), CASSANDRA_THRIFT ("org.janusgraph.diskstorage.cassandra.thrift.CassandraThriftStoreManager", "cassandrathrift"), CASSANDRA_ASTYANAX ("org.janusgraph.diskstorage.cassandra.astyanax.AstyanaxStoreManager", ImmutableList.of ("cassandra", "astyanax"), CASSANDRA_EMBEDDED ("org.janusgraph.diskstorage.cassandra.embedded.CassandraEmbeddedStoreManager", "embeddedcassandra"), CQL ("org.janusgraph.diskstorage.cql.CQLStoreManager", "cql") HBASE ("org.janusgraph.diskstorage.hbase.HBaseStoreManager", "hbase"), IN_MEMORY ("org.janusgraph.diskstorage.keycolumnvalue.inmemory.InMemoryStoreManager", "inmemory"), SEQUOIADB ("org.janusgraph.diskstorage.sequoiadb.SequoiadbStoreManager", "sequoiadb")

The StandardStoreManager class is a module that belongs to janusgraph-cord, so after subsequent compilation, you should replace the old janusgraph-core jar package with the latest jar package.

Then you just download the SequoiaDB for JanusGraph project from github, put it in the root directory of the JanusGraph source code, modify the pom.xml compilation script of maven, add the driver version note of SequoiaDB and add the project compilation of SequoiaDB for JanusGraph.

1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 4.4.1 2.7.7 1.4.10 2.1.5 1.4.10 3.2.1. Janusgraph-doc janusgraph-solr janusgraph-examples janusgraph-sequoiadb

Recompile the JanusGraph:

Finally, save the compiled API driver jar package for janusgraph-sequoiadb-0.4.0.jar and SequoiaDB to the ${JANUSGRAPH_BINARY_HOME} / lib directory, and you are done with JanusGraph extending SequoiaDB storage.

03

JanusGraph configures SequoiaDB as storage

We put the jar package compiled by the module of janusgraph-sequoiadb and the JSON API jar package of SequoiaDB into the lib directory of JunasGraph, and update the jar package of janusgraph-core at the same time, you will complete the operation of JanusGraph extending SequoiaDB storage.

In order for JanusGraph to know the connection information of SequoiaDB, you also need to prepare a config file. Add a janusgraph-sequoiadb.properties file to the conf directory of JunasGraph, which is roughly as follows:

Gremlin.graph=org.janusgraph.core.JanusGraphFactorystorage.backend=sequoiadbstorage.hostname=10.211.55.7storage.port=11810#storage.username=sdbadmin#storage.password=sdbadminstorage.meta.visibility = truecache.db-cache = falsecache.db-cache-clean-wait = 20cache.db-cache-time = 180000cache.db-cache-size = 0.5

Let me introduce you to some important parameters.

The IP address of the storage.hostname,SequoiaDB coord node, or hostname

Port number of the storage.port,SequoiaDB coord node

Storage.username. If SequoiaDB is configured with authentication, then the user name of authentication needs to be configured.

Storage.password. If SequoiaDB is configured with authentication, then you need to configure the password for authentication.

When we have written the configuration information of SequoiaDB into the config file, it is also a smooth thing for JanusGraph to interface with SequoiaDB.

You can directly open the Gremlin control interface, and then just use JanusGraph as usual.

Here, I have prepared a little demo for you.

Graph = JanusGraphFactory.open ('conf/janusgraph-sequoiadb.properties'); graph.addVertex ("name", "aaa", "num", 123) g = graph.traversal () g.V (). Values (' name') the answers to the practical questions on how to carry out SequoiaDB + JanusGraph are shared here. I hope the above content can be of some help to everyone. If you still have a lot of doubts to be solved, you can follow the industry information channel for more related knowledge.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.