Cloning pro-actively for RHEL7.9.z as, although it doesn't look like RHEL7 has the same issue with NVMe devices, we don't know if there are devices out there that would report a too-small optimal i/o size either. A 512B block size can majorly affect gfs2's performance.
+++ This bug was initially created as a clone of Bug #1839219 +++
Description of problem:
When making a GFS2 file system without setting the file system block size, the default GFS2 block size of 4096 is expected. This can be reduced to a smaller size if the device gives a topology with a small optimal_io_size.
We found this was the case with the NMVe devices exported by AWS.
[root@nodea ~]# mkfs.gfs2 -D -O -j4 -p lock_dlm -t bperkins200515w8:brawl1 /dev/mapper/brawl-brawl1
alignment_offset: 0
logical_sector_size: 512
minimum_io_size: 512
optimal_io_size: 512
physical_sector_size: 512
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory (gfs2-utils bug fix and enhancement update), and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHBA-2020:5028