aggregated file inconsitencies when using IOR 4.0.0

[denisbertini]

I try to use IOR with bbview cahing but actually i got inconsistency in the final file size that i do not really understand...

IOR-4.0.0: MPI Coordinated Test of Parallel I/O
Began               : Wed May  6 15:34:00 2026
Command line        : /idril/rz/dbertini/io-tests/iot2//ior/bin/ior -a MPIIO --mpiio.useFileView -w -c -C -e -g -i 1 -vv -s 2000 -b 1m -t 1m -k -o testdir/file
Machine             : Linux ccexe0327 5.14.0-611.49.1.el9_7.x86_64 #1 SMP PREEMPT_DYNAMIC Thu Apr 23 13:13:41 UTC 2026 x86_64
TestID              : 0
StartTime           : Wed May  6 15:34:00 2026
Path                : testdir/file
FS                  : 6979.4 TiB   Used FS: 75.0%   Inodes: 1781.9 Mi   Used Inodes: 2.2%
Participating tasks : 200
Using reorderTasks '-C' (useful to avoid read cache in client)

Options: 
api                 : MPIIO
apiVersion          : (3.1)
test filename       : testdir/file
access              : single-shared-file
type                : collective
segments            : 2000
ordering in a file  : sequential
ordering inter file : constant task offset
task offset         : 1
nodes               : 2
tasks               : 200
clients per node    : 100
repetitions         : 1
xfersize            : 1 MiB
blocksize           : 1 MiB
aggregate filesize  : 390.62 GiB
verbose             : 2

Results: 
Using Time Stamp 1778074440 (0x69fb4348) for Data Signature

access    bw(MiB/s)  IOPS       Latency(s)  block(KiB) xfer(KiB)  open(s)    wr/rd(s)   close(s)   total(s)   iter
------    ---------  ----       ----------  ---------- ---------  --------   --------   --------   --------   ----
Commencing write performance test: Wed May  6 15:34:12 2026
WARNING: inconsistent file size by different tasks
WARNING: Expected aggregate file size       = 419430400000
WARNING: Stat() of aggregate file size      = 59191066624
WARNING: Using actual aggregate bytes moved = 419430400000
write     3794       4109       97.27       1024.00    1024.00    7.45       97.35      0.613807   105.44     0   
Max Write: 3793.81 MiB/sec (3978.09 MB/sec)

I controlled that the size of individual size is at the end 2GB per rank as it should but somehow the FLUSHING process incomplete and the final file size is 100G instead of 400G.
As expected i got 200 calls like:

Flushing file testdir/file for view index 1

in the logs which corresponds to the number of the tasks but somehow it does not aggregate to the proper file size at the end
Is it linked to the flags used in IOR?

 -a MPIIO --mpiio.useFileView -w -c -C -e -g -i 1 -vv -s 2000 -b 1m -t 1m -k -o testdir/file

Any idea?
Best regards
Denis