Pinging a loopback and getting 40% success is a unicast issue. Troubleshoot just like any unicast routing issue. Find the path that the ping takes and examines where in the path the drop is occuring. That will mean pinging each next hop to the destination and until you find the drop. Take the following example:
r1 e1----e2 r2 e3----e4 r4 e5---e6 r6
Loopback on r6 is not pingable or losing packet if ping from r1. from r1 ping the ip address of e3, e4 then e5, then e6. see which of the interfaces is also dropping. This is simplified of course but the point is there is a path to the destination and isolating where the drop along the path, if any.
I hope this helps.
Please rate helpful posts.
