Hi,
I have an old NFS sharing NAS which is performing good, since years, not excellent but good for its job.
Now I have a new NFS sharing NAS which is really slow...
old NAS specs: qnap ts809-rp, 8x 1.5tb SAMSUNG HD154UI 1AG0 in 1 raid5 vol, Core 2 Duo 2.8 GHz, 2GB DDRII RAM, 1GB eth MTU 1500
new NAS specs: qnap ts879-rp, 7x 2.0tb Hitachi HUA723020ALA640 MK7O in 1 raid5 vol, Quad Core Xeon® E3-1225 v2 Processor 3.2 GHz, 4 GB DDR3 ECC RAM, 1GB eth MTU 1500
i've run pveperf test on both, from two cluster nodes, and got these results:
this puzzled me a bit.. I know these are sw raid units, but I expected the new NAS to perform far better than the old...
so I looked at mount options (I mounted those units through pve gui)
I think rsize and wsize are involved... but how can I change them (and to which values, how to find)?
and why pve mounted those in such different ways?
if it can make a difference:
- ts809 was mounted in cluster on a node with old, 2.x before upgrade it to 3.1,
- ts879 was mounted in cluster on another node 2.x, after upgrading it to 3.1
Have you any hints of what is happening and how to use at best the two units? other tests?
I can run tests from cli on both NAS, but that is a custom qnap linux, there's no apt-get, and I don't want to mess something there, if possibile...
Thanks, Marco
I have an old NFS sharing NAS which is performing good, since years, not excellent but good for its job.
Now I have a new NFS sharing NAS which is really slow...
old NAS specs: qnap ts809-rp, 8x 1.5tb SAMSUNG HD154UI 1AG0 in 1 raid5 vol, Core 2 Duo 2.8 GHz, 2GB DDRII RAM, 1GB eth MTU 1500
new NAS specs: qnap ts879-rp, 7x 2.0tb Hitachi HUA723020ALA640 MK7O in 1 raid5 vol, Quad Core Xeon® E3-1225 v2 Processor 3.2 GHz, 4 GB DDR3 ECC RAM, 1GB eth MTU 1500
i've run pveperf test on both, from two cluster nodes, and got these results:
Code:
# pveperf /mnt/pve/ts809/ (old NAS)
CPU BOGOMIPS: 72527.28
REGEX/SECOND: 947321
HD SIZE: 9617.06 GB (ts809:/myshare)
FSYNCS/SECOND: 1740.50
DNS EXT: 57.34 ms
DNS INT: 1.33 ms
# pveperf /mnt/pve/ts879/ (new NAS)
CPU BOGOMIPS: 72527.28
REGEX/SECOND: 948892
HD SIZE: 11081.12 GB (ts879:/myshare)
FSYNCS/SECOND: 204.80
DNS EXT: 135.45 ms
DNS INT: 0.67 msso I looked at mount options (I mounted those units through pve gui)
Code:
IP_of_ts809:/iso_images on /mnt/pve/ts809 type nfs (rw
relatime
vers=3
rsize=262144
wsize=262144
namlen=255
hard
proto=tcp
timeo=600
retrans=2
sec=sys
mountaddr=IP_of_ts809
mountvers=3
mountport=659
mountproto=udp
local_lock=none
addr=IP_of_ts809)
IP_of_ts879:/Download on /mnt/pve/ts879 type nfs (rw
relatime
vers=3
rsize=32768
wsize=32768
namlen=255
hard
proto=tcp
timeo=600
retrans=2
sec=sys
mountaddr=IP_of_ts879
mountvers=3
mountport=59669
mountproto=udp
local_lock=none
addr=IP_of_ts879)and why pve mounted those in such different ways?
if it can make a difference:
- ts809 was mounted in cluster on a node with old, 2.x before upgrade it to 3.1,
- ts879 was mounted in cluster on another node 2.x, after upgrading it to 3.1
Have you any hints of what is happening and how to use at best the two units? other tests?
I can run tests from cli on both NAS, but that is a custom qnap linux, there's no apt-get, and I don't want to mess something there, if possibile...
Thanks, Marco