The client is VMWare ESXi, so my options are limited. I tried enabling jumbo frames (used 9000) and this made very little difference.
-----Original Message----- From: Zé Loff <zel...@zeloff.org> Sent: Tuesday, December 5, 2023 10:12 AM To: Steven Surdock <ssurd...@engineered-net.com> Cc: misc@openbsd.org Subject: Re: NFS Server performance On Tue, Dec 05, 2023 at 02:06:44PM +0000, Steven Surdock wrote: > Using an OBSD 7.4 VM on VMware as an NFS server on HOST02. It is primarily > used to store VMWare VM backups from HOST01, so VMWare is the NFS client. > I'm seeing transfers of about 1.2 MB/s. > > SCP from HOST01 to OBSD VM (same filesystem) copies at 110 MB/s. > Iperf3 from a VM on HOST01 to OBSD on HOST02 gives me 900+ mbps. > OBSD is a stock install running -stable. > NFS is using v3 (according to VMWare) and using TCP During the NFS > transfer the RECV-Q on the OBSD interface runs either 64000+ or 0. > I tried both em and vmx interface types. > > /etc/rc.conf.local: > mountd_flags="" # for normal use: "" > nfsd_flags="-tun 4" # Crank the 4 for a busy NFS fileserver > ntpd_flags="" # enabled during install > portmap_flags="" # for normal use: "" > > Any clues on where to look to (greatly) improve NFS performance would be > appreciated. Increasing write size, read size and the read-ahead count on the client has helped me. E.g., on the client's fstab: 10.17.18.10:/shared/stuff /nfs/stuff nfs rw,nodev,nosuid,intr,tcp,bg,noatime,-a=4,-r=32768,-w=32768 0 0 Cheers Zé --
