We just installed Sense 3.1 SR3 on a scheduler rim node to manage reloads for a central node (central node's scheduler only acts as master). Central and rim node are installed on two separate machines, same version of OS and Qlik Sense.
We have created a sync rule to sync all apps with between the two nodes wich worked fine for a few days. But now the binary sync has stop to work, and is we create a new app (on the central node) and try to reload it we get a error becouse of missing slave node. And if we look at the rim nodes cpu usage the repository service keeps hitting the cpu with 30-60% usage.
It's seems to me that the syncronisation between the nodes stops working after a few days (it starts to work if we restart both machines), but the repository service are doing something demanding due to the cpu usage.