In large-scale routing scenarios，enabled inline-jflow on MPC4E may reports "Sample RR (DFW) does not have enough room" error message.

Refer below logs:

Apr 30 00:54:54 mx fpc0 jnh_partition_ok_per_inst(5054): jnh_expand_partition failed, inst 0, jnh_app Inline Jflow Sample RR (DFW), dwords 65536

Apr 30 00:54:54 mx fpc0 jnh_add_prefix_table(474): PFE 255 Table Partition, Inline Jflow Sample RR (DFW) does not have enough room

If expand memory is exhausted, there is no room to install sample-rr routes. 

On production, if the device route scale has reached about 2.5 million, it could be scaling limit.

Using below command to check pfe0(jnh0) or pfe1(jnh1) usage, if EDMEM usage is full that means expand memory will be used.

user@mx> request pfe execute command "show jnh 0 pool usage bytes" target fpc0
EDMEM overall usage:
[NH////////////|FW//////////////////|CNTR///////////|HASH//////////////|ENCAPS////|]
0       48.0         114.7      162.7       223.2   255.7 256.0MB

Then using below command to check pfe0(jnh0) or pfe1(jnh1) memory detail info. 

If "Next Hop" shared memory has no free bytes, the pfe will reports this error message:

user@mx> request pfe execute command "show jnh 0 pool detail bytes" target fpc0
        Name   PhysAddr      Size (bytes) by Type        Used(b)  (%)  Free(b)  (%)
                      Private  Shared   Reserved  
      Next Hop   0 0x00000000    33553408     0     0   29680504 88%   3872904 12%
                 1 0x00e00000       0  33553408     0   33521592 99%    31816  1%
                 3 0xc0000000     393216  2129920     0    2465784 97%    57352  3%

This is scaling limit issue, to resolve this issue, need to decrease route scale on device, or use other type MPC with more memory.