The Hbase is an open source distributed database on Hadoop, it is robust regardless of the performance problem. The ObjectDB cluster may be a distributed database, I guess. But I dont know whether it has only one entry, or each node has an entry.
Could you please explain the difference between Hbase and ObjectDB clustering?
TIA
ObjectDB supports master-slave cluster, in which WRITE requests are served by one master server and READ requests can be served by all the servers. All the servers in an ObjectDB cluster maintain a full copy of the database.
In HBase data is divided between nodes. All the servers handle WRITE requests and READ requests of data that they manage. This is usually slower but it provides better scalability when the data is too large for one server and when there is a need for more than one server to handle WRITE requests.
As you said, the ObjectDB clustering can not support a database whose size is beyond the size of the master node, right? It is better to handle scalable data for ObjectDB clustering.
BTW, how to define a priory order for visiting a slave than a master? For the reading balance, one may want to access female information by the master node, and access male information by the slave-1 node, and access other information by the slave-2 node. Only when the defined node failed, the secondary node should provider service.
TIA
Given one master and two slaves defined, how to add or delete a slave to change the redundancy of the clustering without restarting.
TIA
Future ObjectDB versions may store data on top of the Hadoop distributed file system (as HBase does) to support larger databases, but that will come with a price tag of slower performance.
In using composite urls the order of the urls defines the connection priority.
No need to restart the master server to add a new slave since the configuration of the master-slave connection is on the slave side only.
>but that will come with a price tag of slower performance.
Yes, HDFS can not support rand write access. I can not imagine how do you maintain a flexible database index.
Eagerly, I want to test your working on Hadoop.
TIA
Hi, what status about the distributed version?
Integration with Hadoop distributed file system is still at the idea level.
Then, how about the support to multi-core machines?
Multi-core machines have always been supported.
Hi, buddy. How about the Objectdb on Hadoop now?