Scheduler limits...
Since 2 days I'm trying to use the scheduler, as I started 16 new projects (other are coming), and my last new VPS is not so able to work well with 44 projects in total.
Using the SCHEDULER, I can decide only which projects to run and the amount of minutes (or also there is the option of the exactly starting time and stop, but I don't need it now).
Later, SER decide which project to start first and which to start on the second start. I think it can be better if I can decide which project to start first and which later.
My example (if it's not clear what I writed).
I want to use the scheduler for my 44 projects (11 websites, 4 tiers each one).
If I set the scheduler to run every 6 hours with 11 projects every time, I select all the projects, later set the minutes (360) and the amount of projects to start everytime (11). When I push start, SER randomly start 11 projects, and after 6 hours stop them and start other random 11 among the restant 33 projects, and so on.
The problem is that, for example, the T1 projects need less time and resources to finish, instead the T3, T2 and T2S need more time because they have to build much more links than the T1s. So, if for example SER start a random group with mayority of T3 and T2, and later a group with mayority of T1, these groups use very different amount of resources (ram and cpu), and sometime the VPS is over the limit, sometime is under its limits.
If I can decide which projects to start, I can manage groups of projects more equilibrate to run together.
Hope my question is clear, thanks for any suggestion!
Using the SCHEDULER, I can decide only which projects to run and the amount of minutes (or also there is the option of the exactly starting time and stop, but I don't need it now).
Later, SER decide which project to start first and which to start on the second start. I think it can be better if I can decide which project to start first and which later.
My example (if it's not clear what I writed).
I want to use the scheduler for my 44 projects (11 websites, 4 tiers each one).
If I set the scheduler to run every 6 hours with 11 projects every time, I select all the projects, later set the minutes (360) and the amount of projects to start everytime (11). When I push start, SER randomly start 11 projects, and after 6 hours stop them and start other random 11 among the restant 33 projects, and so on.
The problem is that, for example, the T1 projects need less time and resources to finish, instead the T3, T2 and T2S need more time because they have to build much more links than the T1s. So, if for example SER start a random group with mayority of T3 and T2, and later a group with mayority of T1, these groups use very different amount of resources (ram and cpu), and sometime the VPS is over the limit, sometime is under its limits.
If I can decide which projects to start, I can manage groups of projects more equilibrate to run together.
Hope my question is clear, thanks for any suggestion!
Comments
However, I think a better solution would be for the scheduler to ignore active(p) projects, although that's more work for @sven...
Actually, when a project is on: active(P), is not consuming any resource? I always wonder if, when some projects are on active(P), I can save resources stopping them for a while, or it doesn't change nothing leaving them active....