we have a MX240 with a RE-S-2000 which is EOL an only has one CPU.
The IPv4 fulltable currently has around 730k routes and after an RE restart and waiting for 3 hours, the FIB only has ~190k routes. When looking with "top" in "start shell" I see that the `rpd` process takes around 50-60% CPU and is in state "kqread" - I think I saw it alot of times in state "active" after a commit which affected route updates in the FIB.
Maybe something changed in Junos 17.3?
We recently updated to 17.3 from 15.1.
Question: Is this slow route installation caused by the old/EOL RE / would a faster RE speed things up?