I want to bring up the idea of solely support PostgreSQL for DHIS 2 for production use to get feedback from the community.
The DHIS 2 team is seeing clear benefits of moving to PostgreSQL-only. The main ones are:
Supporting and testing for more than one database platform is time-consuming. The majority of DHIS 2 implementations is using PostgreSQL already.
PostgreSQL supports advanced geospatial features through PostGIS. We are already depending on these for the GIS module to work optimally (i.e. for event clustering in GIS). There are no direct equivalent in e.g. MySQL. We are also considering using PostGIS data types in main data model.
PostgreSQL will provide some impressive features which will become very useful for us in version 10 and 11. E.g. native partitioning, parallel queries, logical replication, foreign tables and improved query performance will be useful and allow us to develop the next DHIS 2 analytics solution.
We are aware this will be inconvenient for users of other databases such as MySQL, but there is the option of migrating your database through the DHIS 2 metadata and data value import-export feature.
On a technical note, we will keep the Hibernate development framework and keep supporting the H2 database for unit testing purposes.
Hi Lars,
Make a lot of sense, and would free time in your team to provide more value (seems you are not lacking ideas). I would say go for it, with maybe dedicated documentation about how to move (even if it’s basically using the import/export feature “as is”, may be worth a couple of hours to write a step by step guide).
Martin
···
On Fri, Sep 8, 2017 at 10:26 AM, Lars Helge Øverland lars@dhis2.org wrote:
Hi all,
I want to bring up the idea of solely support PostgreSQL for DHIS 2 for production use to get feedback from the community.
The DHIS 2 team is seeing clear benefits of moving to PostgreSQL-only. The main ones are:
Supporting and testing for more than one database platform is time-consuming. The majority of DHIS 2 implementations is using PostgreSQL already.
PostgreSQL supports advanced geospatial features through PostGIS. We are already depending on these for the GIS module to work optimally (i.e. for event clustering in GIS). There are no direct equivalent in e.g. MySQL. We are also considering using PostGIS data types in main data model.
PostgreSQL will provide some impressive features which will become very useful for us in version 10 and 11. E.g. native partitioning, parallel queries, logical replication, foreign tables and improved query performance will be useful and allow us to develop the next DHIS 2 analytics solution.
We are aware this will be inconvenient for users of other databases such as MySQL, but there is the option of migrating your database through the DHIS 2 metadata and data value import-export feature.
On a technical note, we will keep the Hibernate development framework and keep supporting the H2 database for unit testing purposes.
Hi Lars,
I think like the DHIS2 is a system that needs to be deployed on a large scale, the choice to keep postgresql is ideal. In its recent version Postgresql becomes more flexible when there is a lot of data. It becomes very close to Oracle. the possibility of using the Json and its highly developed GIS module is a major asset. I’m ok to keep Postgresql as DHIS2 DBMS.
I fully agree with your assessment and the benefits are clear.
Best regards,
JM
El 08/09/2017, a las 10:26, Lars Helge Øverland lars@dhis2.org escribió:
Hi all,
I want to bring up the idea of solely support PostgreSQL for DHIS 2 for production use to get feedback from the community.
The DHIS 2 team is seeing clear benefits of moving to PostgreSQL-only. The main ones are:
Supporting and testing for more than one database platform is time-consuming. The majority of DHIS 2 implementations is using PostgreSQL already.
PostgreSQL supports advanced geospatial features through PostGIS. We are already depending on these for the GIS module to work optimally (i.e. for event clustering in GIS). There are no direct equivalent in e.g. MySQL. We are also considering using PostGIS data types in main data model.
PostgreSQL will provide some impressive features which will become very useful for us in version 10 and 11. E.g. native partitioning, parallel queries, logical replication, foreign tables and improved query performance will be useful and allow us to develop the next DHIS 2 analytics solution.
We are aware this will be inconvenient for users of other databases such as MySQL, but there is the option of migrating your database through the DHIS 2 metadata and data value import-export feature.
On a technical note, we will keep the Hibernate development framework and keep supporting the H2 database for unit testing purposes.
Agree. As a prelude we can put the export import through its paces. Would be great to have a volunteer mysql Guinea pig user that we could document a migration.
I want to bring up the idea of solely support PostgreSQL for DHIS 2 for production use to get feedback from the community.
The DHIS 2 team is seeing clear benefits of moving to PostgreSQL-only. The main ones are:
Supporting and testing for more than one database platform is time-consuming. The majority of DHIS 2 implementations is using PostgreSQL already.
PostgreSQL supports advanced geospatial features through PostGIS. We are already depending on these for the GIS module to work optimally (i.e. for event clustering in GIS). There are no direct equivalent in e.g. MySQL. We are also considering using PostGIS data types in main data model.
PostgreSQL will provide some impressive features which will become very useful for us in version 10 and 11. E.g. native partitioning, parallel queries, logical replication, foreign tables and improved query performance will be useful and allow us to develop the next DHIS 2 analytics solution.
We are aware this will be inconvenient for users of other databases such as MySQL, but there is the option of migrating your database through the DHIS 2 metadata and data value import-export feature.
On a technical note, we will keep the Hibernate development framework and keep supporting the H2 database for unit testing purposes.
Fully agreed. To concentrate one one DB platforms will not only saves time and effort, also being efficiency to the system and can make more use of advance functionality of the PostgreSQL & PostGIS.
You will have our full support.
Hannan
···
On Fri, Sep 8, 2017 at 2:26 PM, Lars Helge Øverland lars@dhis2.org wrote:
Hi all,
I want to bring up the idea of solely support PostgreSQL for DHIS 2 for production use to get feedback from the community.
The DHIS 2 team is seeing clear benefits of moving to PostgreSQL-only. The main ones are:
Supporting and testing for more than one database platform is time-consuming. The majority of DHIS 2 implementations is using PostgreSQL already.
PostgreSQL supports advanced geospatial features through PostGIS. We are already depending on these for the GIS module to work optimally (i.e. for event clustering in GIS). There are no direct equivalent in e.g. MySQL. We are also considering using PostGIS data types in main data model.
PostgreSQL will provide some impressive features which will become very useful for us in version 10 and 11. E.g. native partitioning, parallel queries, logical replication, foreign tables and improved query performance will be useful and allow us to develop the next DHIS 2 analytics solution.
We are aware this will be inconvenient for users of other databases such as MySQL, but there is the option of migrating your database through the DHIS 2 metadata and data value import-export feature.
On a technical note, we will keep the Hibernate development framework and keep supporting the H2 database for unit testing purposes.
On Fri, Sep 8, 2017 at 1:19 PM, Hannan Khan hannank@gmail.com wrote:
Hi Lars
Fully agreed. To concentrate one one DB platforms will not only saves time and effort, also being efficiency to the system and can make more use of advance functionality of the PostgreSQL & PostGIS.
You will have our full support.
Hannan
–
On Fri, Sep 8, 2017 at 2:26 PM, Lars Helge Øverland lars@dhis2.org wrote:
Hi all,
I want to bring up the idea of solely support PostgreSQL for DHIS 2 for production use to get feedback from the community.
The DHIS 2 team is seeing clear benefits of moving to PostgreSQL-only. The main ones are:
Supporting and testing for more than one database platform is time-consuming. The majority of DHIS 2 implementations is using PostgreSQL already.
PostgreSQL supports advanced geospatial features through PostGIS. We are already depending on these for the GIS module to work optimally (i.e. for event clustering in GIS). There are no direct equivalent in e.g. MySQL. We are also considering using PostGIS data types in main data model.
PostgreSQL will provide some impressive features which will become very useful for us in version 10 and 11. E.g. native partitioning, parallel queries, logical replication, foreign tables and improved query performance will be useful and allow us to develop the next DHIS 2 analytics solution.
We are aware this will be inconvenient for users of other databases such as MySQL, but there is the option of migrating your database through the DHIS 2 metadata and data value import-export feature.
On a technical note, we will keep the Hibernate development framework and keep supporting the H2 database for unit testing purposes.