Discovery via a shared database instead of multicast or seed IPs

classic Classic list List threaded Threaded
2 messages Options
dsetrakyan dsetrakyan
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Discovery via a shared database instead of multicast or seed IPs

Many of our applications are 5-15 nodes, not 100s. They have shared databases which are highly available and our customers are well versed at their care and feeding. Looks like the out of the box options for Ignite discovery are multicast, seed IPs, or EC2 (not applicable). I was hoping to find a JDBC discovery mechanism as well. Since we always have an HA database that is a "shared" artifact in our architecture, it would be nice to remove the configuration of Ignite discovery and just let it use some table with the IPs, ports, and heartbeats of all of the nodes in the topology (or at least some seed ones). Then I would only have to configure the database URL (which I'm already doing) and wouldn't need to complicate things further with IP addresses or multicast.

Is this possible today? I see TcpDiscoveryJdbcIpFinder, is that what I'm looking for? Anyone have an example of how to use it?

----------
This post is migrated from Apache Ignite forum on readme.io (no longer active)
dsetrakyan dsetrakyan
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Discovery via a shared database instead of multicast or seed IPs

Yes, it is supported and you can use TcpDiscoveryJdbcIpFinder to configure it. Looks like the only configuration parameter it requires is the DataSource (setDataSource(...) method).

It is also documented in JDBC Based Discovery section (you can find example there).

D.

On Fri, May 8, 2015 at 12:42 AM, Dmitriy Setrakyan <[hidden email]> wrote:
Many of our applications are 5-15 nodes, not 100s. They have shared databases which are highly available and our customers are well versed at their care and feeding. Looks like the out of the box options for Ignite discovery are multicast, seed IPs, or EC2 (not applicable). I was hoping to find a JDBC discovery mechanism as well. Since we always have an HA database that is a "shared" artifact in our architecture, it would be nice to remove the configuration of Ignite discovery and just let it use some table with the IPs, ports, and heartbeats of all of the nodes in the topology (or at least some seed ones). Then I would only have to configure the database URL (which I'm already doing) and wouldn't need to complicate things further with IP addresses or multicast.

Is this possible today? I see TcpDiscoveryJdbcIpFinder, is that what I'm looking for? Anyone have an example of how to use it?

----------
This post is migrated from Apache Ignite forum on readme.io (no longer active)

Loading...