Managing a big repository

Classic Servoy
1

Just thinking: I use Servoy before version 1.0 and I think I’ll continue doing it for a long time :-)

That said, what about a growing repository, with dozen of solutions and a lot of database to open?
I noticed that starting up a fresh new Servoy installation is definitely faster than my working copy (and I’m keeping only the solutions in production, not all of them).

What’s your approach about that?

2

It depends on what database you are using for the backend. Some databases require periodic maintenance (think of PostgreSQL’s vacuumdb) to stop them from becoming slow.

Note that in a production environment Servoy performance should not degrade with a higher number of solutions (unless your database performance degrades due to reasons outside of Servoy itself); Servoy Application Server loads all active solutions from your repository on startup into memory, so database access to your repository is virtually eliminated.

In a developer environment it is possible for performance to become a bit slower with many solutions in the database. However if you have a good database this should not be a significant performance loss.

3

Hi Ric,

I decided to set up repositories for different purposes.

I only start up one as the main repository and have as many as 4 possible instances of developer on 2-4 different repositories.

One remote for one of my customers, the same but local, one for several customers and my own development/test stuff as the main repository.

Works like a charm for me and definitely keeps startup times within the limit I set for myself (or am I just getting used to the startup :) )

4

sebster:
It depends on what database you are using for the backend. Some databases require periodic maintenance (think of PostgreSQL’s vacuumdb) to stop them from becoming slow.

I use Sybase.
But my problem is that a repository with 50 solutions and, say, 30 different db’s to open in database directory will probably slow down the startup, since I’m already seeing some difference with 8 db and 15 solutions.

sebster:
In a developer environment it is possible for performance to become a bit slower with many solutions in the database. However if you have a good database this should not be a significant performance loss.

It’s not really a performance problem, but a startup problem what I was considering.

5

IT2BE:
Hi Ric,

I decided to set up repositories for different purposes.

I only start up one as the main repository and have as many as 4 possible instances of developer on 2-4 different repositories.

One remote for one of my customers, the same but local, one for several customers and my own development/test stuff as the main repository.

Works like a charm for me and definitely keeps startup times within the limit I set for myself (or am I just getting used to the startup :) )

It’s an option :-)

Thanks, Marcel