Scheduler running on same server as iceccd

[Nik Twerdochlib]

I have a collection of 64 cpu servers configured and one of these systems also runs the scheduler.  Everything works well except that I have noticed that the system hosting the scheduler will rarely accept jobs and when it does it seems to never be more then 2-4.

 

Been searching through documentation, google searches, and the code but I have not found anything stating this could be an issue.  Anyone have any thoughts or have also seen this behavior?

I am running icecream 1.3.1. The servers are gentoo 2.6 Kernel 4.19.97

[Nik Twerdochlib]

I have found that I might be wrong about this.  Switch to using icecream-sundae on the scheduler server does show an average number of jobs being run on all nodes.  It was the gui monitor that was giving me the impression the scheduler node was being avoided, which could be due to many things other then what I interpreted it to be.