I remember an issue like this with early shipments of RHEL7. I recall it as a defect in kNFS, My memory could be wrong though. These are exchanged very early on, so I imagine this is due to NFS client (Centos 7.2). Try with Centos 7.6, if you don't see this issue then it is clear where the issue is.

On Wed, Sep 25, 2019 at 9:03 AM QR <zhbingyin@sina.com> wrote:
In CentOS7.2 under v4.0, the wsize is 1048576, but when writing 1048576 bytes on a file, ganesha receives 4 requests.
The problem is who splits the 1MB request to 4 requests(256KB)?
Is anyone know about this? Thanks in advance.

Ganesha server info
    ganesha version: V2.7.6
    FSAL           : gluster
nfs client info
    nfs version    : nfs v4.0
    client info    : CentOS Linux release 7.2.1511(3.10.0-327.el7.x86_64)
_______________________________________________
Devel mailing list -- devel@lists.nfs-ganesha.org
To unsubscribe send an email to devel-leave@lists.nfs-ganesha.org