Remove the MPI_Barrier after big_file_mpi_close

We have been told by Frontera admin that it is causing problems for
Astrid, as multiple files are written in quick succession and this
forces the file servers to sync, rather than rebalancing to a different
server.

src/bigfile-mpi.c

@@ -106,7 +106,7 @@ big_file_mpi_create_block(BigFile * bf,
     size_t fsize[Nfile];
     int i;
     for(i = 0; i < Nfile; i ++) {
-        fsize[i] = size * (i + 1) / Nfile 
+        fsize[i] = size * (i + 1) / Nfile
                  - size * (i) / Nfile;
     }
     return _big_file_mpi_create_block(bf, block, blockname, dtype,
@@ -144,7 +144,6 @@ _big_file_mpi_create_block(BigFile * bf,
 int big_file_mpi_close(BigFile * bf, MPI_Comm comm) {
     if(comm == MPI_COMM_NULL) return 0;
     int rt = big_file_close(bf);
-    MPI_Barrier(comm);
     return rt;
 }
 
@@ -155,7 +154,7 @@ _big_block_mpi_open(BigBlock * bb, const char * basename, MPI_Comm comm)
     int rank;
     MPI_Comm_rank(comm, &rank);
     int rt;
-    if(rank == 0) { 
+    if(rank == 0) {
         rt = _big_block_open(bb, basename);
     } else {
         rt = 0;
@@ -443,7 +442,7 @@ _throttle_action(MPI_Comm comm, int concurrency, BigBlock * block,
         if(0 != (rt = big_file_mpi_broadcast_anyerror(rt, seggrp->Group))) {
             /* failed , abort. */
             continue;
-        } 
+        }
         if(seggrp->ThisSegment != segment) continue;
 
         /* use the offset on the first task in the SegGroup */