The explicit-null statement enables the PE router to send an MPLS label with a value of 0 (explicit null) instead of an IP header for packets destined to the VPN customer sites. This allows the penultimate hop router (the router before the egress PE router) to preserve the EXP bits of the MPLS label and pass them to the egress PE router. The egress PE router can then use these EXP bits to classify the traffic according to the CoS policy2
. In this example, PE-1 should configure the explicit-null statement under [edit protocols mpls label-switched-path PE-1_to_PE-2] hierarchy level.

BGP route reflectors are BGP routers that are allowed to ignore the IBGP loop avoidance rule and advertise IBGP learned routes to other IBGP peers under specific conditions. BGP route reflectors can reduce the number of IBGP sessions and updates in a network by eliminating the need for a full mesh of IBGP peers.
BGP route reflectors can have three types of peerings:
EBGP neighbor: A BGP router that belongs to a different autonomous system (AS) than the route reflector.
IBGP client neighbor: An IBGP router that receives reflected routes from the route reflector. A client does not need to peer with other clients or non-clients.
IBGP non-client neighbor: An IBGP router that does not receive reflected routes from the route reflector. A non-client needs to peer with other non-clients and the route reflector.
In the exhibit, we can see that RR1 and RR2 are route reflectors in the same AS with reachability redundancy.
They have two types of peerings: EBGP neighbors (R1 and R4) and IBGP client neighbors (Client1, Client2, and Client3). RR1 and RR2 are also peered with each other as IBGP non-client neighbors.

the default behavior for iBGP is to propagate EBGP-learned prefixes without changing the next-hop. This can cause issues if the next-hop is not reachable via the IGP. One solution is to use the next-hop self command on R4, which will change the next-hop attribute to its own loopback address. This way, R5 can reach the next-hop via the IGP and install the route in its routing table.
Another solution is to add the external interface prefix (120.0.4.16/30) to the IGP routing tables of R4 and R5.
This will also make the next-hop reachable via the IGP and allow R5 to use the route. According to 2, this is a possible workaround for a pure IP network, but it may not work well for an MPLS network.
The reason why the route is being hidden is that R5 cannot reach the BGP next hop 10.0.0.1, which is the address of R1. R5 does not have a route to 10.0.0.0/24 in its routing table, and neither does R4. Therefore, R5 cannot resolve the BGP next hop and marks the route as hidden.
There are two solutions that will solve this problem:
* Option A: On R4, create a policy to change the BGP next hop to itself and apply it to IBGP as an export policy. This way, R5 will receive the route with a next hop of 172.16.1.2, which is reachable via the IGP. This solution is also known as next-hop-self1.
* Option B: Add the external interface prefix to the IGP routing tables. This way, R4 and R5 will learn a route to 10.0.0.0/24 via the IGP and be able to resolve the BGP next hop. This solution is also known as recursive lookup2.
Option C is not correct because adding the internal interface prefix to the BGP routing tables will not help R5 reach the BGP next hop 10.0.0.1.
Option D is not correct because changing the BGP next hop to 172.16.1.1 on R4 will not help R5 either, since R5 does not have a route to 172.16.1.1 in its routing table.
References: 1: Configuring Next-Hop-Self for IBGP Peers 2: Understanding Recursive Lookup