So I have two workstations with MySQL and two unicenta DB's. I'm trying to simulate a CORP and STORE replication. Now my silly question is this: does the SymmetricDS installation need to be on both workstations? Or can the CORP workstation "poll" the STORE for any new changes to download without the STORE having the SymmetricDS binaries installed?
I know it should be obvious..
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
You can configure it as a multi-homing environment.
Meaning, Symmetric works similar to a 'gateway' and you configure your engines properties files to point to your dbs.
I usually have Symmetric configured in a server that doesn't even has a database ! I configure it on a Linux box and point its engines properties files to my dbs [through local or remote network].
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Now I'm totally confused about group_links specifically P & W.. push/pull. These directives are relative to which node!?
For example:
1. ('CORP', 'STORE', 'P')
2. ('STORE', 'CORP', 'W');
In "1." the CORP is pushing data to STORE? Right? BUT, in "2." who's waiting for a PULL? Is it the STORE is waiting for a PULL or is CORP waiting for the PULL? Is this relative to the first/source node always?
Very confusing..
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The first column is the source node group, the second column is the target node group. In the example above, the CORP node group is configured to PUSH data to the STORE node group, and the STORE node group is configured to wait for pull from the CORP node group. If you are doing a bi-directional sync between the CORP and STORE node groups, this configuration means that the CORP SymmetricDS agent will make a connection to a STORE SymmetricDS agent and push its data changes to the store. The STORE will wait for a connection to be made by the CORP to move its data up to the CORP. This type of PUSH / WAIT FOR PULL configuration allows you to do a bi-directional sync while only poking a hole through a firewall in one direction.
Last edit: Greg Wilmer 2015-04-17
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
So I have two workstations with MySQL and two unicenta DB's. I'm trying to simulate a CORP and STORE replication. Now my silly question is this: does the SymmetricDS installation need to be on both workstations? Or can the CORP workstation "poll" the STORE for any new changes to download without the STORE having the SymmetricDS binaries installed?
I know it should be obvious..
You can configure it as a multi-homing environment.
Meaning, Symmetric works similar to a 'gateway' and you configure your engines properties files to point to your dbs.
I usually have Symmetric configured in a server that doesn't even has a database ! I configure it on a Linux box and point its engines properties files to my dbs [through local or remote network].
Great, thanks!
Now I'm totally confused about group_links specifically P & W.. push/pull. These directives are relative to which node!?
For example:
1. ('CORP', 'STORE', 'P')
2. ('STORE', 'CORP', 'W');
In "1." the CORP is pushing data to STORE? Right? BUT, in "2." who's waiting for a PULL? Is it the STORE is waiting for a PULL or is CORP waiting for the PULL? Is this relative to the first/source node always?
Very confusing..
The first column is the source node group, the second column is the target node group. In the example above, the CORP node group is configured to PUSH data to the STORE node group, and the STORE node group is configured to wait for pull from the CORP node group. If you are doing a bi-directional sync between the CORP and STORE node groups, this configuration means that the CORP SymmetricDS agent will make a connection to a STORE SymmetricDS agent and push its data changes to the store. The STORE will wait for a connection to be made by the CORP to move its data up to the CORP. This type of PUSH / WAIT FOR PULL configuration allows you to do a bi-directional sync while only poking a hole through a firewall in one direction.
Last edit: Greg Wilmer 2015-04-17
Great explanation!! This should go in the manual, it couldn't be more clear. Now I totally get it and understand why it's setup like this.
Thank you!