I have three machines that all have the same hardware configuration. They all serve as nodes in a Proxmox cluster.
Two of them have been running VMs over NFS for months with no issues.
This is on a fresh installation of Proxmox 6.0. On a previous fresh installation of Proxmox 5.4, the
I'm not sure where to go from here. Being that the hardware is the same as the other two machines, and both of those other machines also perform fine on fresh installs, I expect this third machine to reach the >2000 FSYNC/SECOND that the other two consistently receive.
Two of them have been running VMs over NFS for months with no issues.
pveperf
on these, targeting the NFS share, shows ~2000 fsync/sec. Trying to boot off of the third node however, also under the same NFS share, basically doesn't work. Takes forever. Seems to take up 100% of a single CPU (vcore) during the boot process and takes about 13 hours to finally load the OS image from NFS into RAM and boot. Running pveperf
on the NFS share on _this_ machine produces these results:
Code:
CPU BOGOMIPS: 166420.16
REGEX/SECOND: 1696783
HD SIZE: 20446.67 GB (storage.family.site:/mnt/zcore/vm_storage)
FSYNCS/SECOND: 166.79
DNS EXT: 45.01 ms
DNS INT: 27.57 ms (family.site)
This is on a fresh installation of Proxmox 6.0. On a previous fresh installation of Proxmox 5.4, the
FSYNC/SECOND
was usually <1. Yes. Less than 1.I'm not sure where to go from here. Being that the hardware is the same as the other two machines, and both of those other machines also perform fine on fresh installs, I expect this third machine to reach the >2000 FSYNC/SECOND that the other two consistently receive.