ZFS with CIFS dedupe and compress too slow

[stevenfoo]

I have installed the latest version of NAS4Free into our 2 x Quad Core CPU with 8GB of memory.

partition - 1 (400GB x 2 near line SATA - hardware RAID 1) - NAS4Free Installed with UFS filesystem
partition - 2 (400GB x 4 near line SATA - hardware RAID 5) - ZFS filesystem (CIFS/SMB - share).

We are getting a very low performance on the ZFS write when copying files from the window server. Any idea why so slow?

zpool iostat
capacity operations bandwidth
pool alloc free read write read write
------------ ----- ----- ----- ----- ----- -----
ZFSdatastore 143G 1.22T 58 40 237K 2.41M

Note:
ZFS dataset is enable with compression and dedupe (hosting fileserver).
Network = 1GB LAN
SMB config as below:

/var/etc/smb.conf
[global]
encrypt passwords = yes
netbios name = XXXXXX
workgroup = XXXXXX
server string = nas4free
security = ads
max protocol = SMB2
dns proxy = no
# Settings to enhance performance:
strict locking = no
read raw = yes
write raw = yes
oplocks = yes
max xmit = 65535
deadtime = 15
getwd cache = yes
socket options = TCP_NODELAY SO_SNDBUF=64240 SO_RCVBUF=64240
# End of performance section
password server = XXXX.XXXX.com
wins server = XXX.XXX.XXX.XXX
unix charset = UTF-8
store dos attributes = yes
local master = no
domain master = no
preferred master = no
os level = 0
time server = no
guest account = ftp
map to guest = Never
display charset = LOCALE
max log size = 100
syslog only = yes
syslog = 1
load printers = no
printing = bsd
printcap name = /dev/null
disable spoolss = yes
log level = 1
dos charset = ASCII
smb passwd file = /var/etc/private/smbpasswd
private dir = /var/etc/private
passdb backend = tdbsam
allow trusted domains = yes
idmap config * : backend = tdb
idmap config * : range = 10000-39999
idmap config XXXXXX : backend = rid
idmap config XXXXXX : range = 10000-39999
winbind enum users = yes
winbind enum groups = yes
winbind use default domain = yes
winbind normalize names = yes
template homedir = /mnt
template shell = /bin/sh
aio read size = 4096
aio write size = 4096

[CIFSShare]
comment = CIFSShare
path = /ZFSdatastore/ZFSdataset
writeable = yes
printable = no
veto files = /.snap/.sujournal/
hide dot files = yes
guest ok = yes
inherit permissions = yes
vfs objects = shadow_copy2 zfsacl
shadow:format = auto-%Y%m%d-%H%M%S
shadow:snapdir = .zfs/snapshot
shadow:sort = desc
shadow:localtime = yes
veto files = /.zfs/

ZFS dataset settings
zfs get -s local all
NAME PROPERTY VALUE SOURCE
ZFSdatastore mountpoint /ZFSdatastore local
ZFSdatastore/ZFSdataset quota none local
ZFSdatastore/ZFSdataset reservation none local
ZFSdatastore/ZFSdataset compression on local
ZFSdatastore/ZFSdataset atime on local
ZFSdatastore/ZFSdataset readonly off local
ZFSdatastore/ZFSdataset snapdir hidden local
ZFSdatastore/ZFSdataset canmount on local
ZFSdatastore/ZFSdataset dedup on local
ZFSdatastore/ZFSdataset sync standard local

[ku-gew]

Probably because you enabled dedup without making careful evaluations of its usefulness. How much is it saving? dedup is usually not worth unless you have savings of 3-4x.

[stevenfoo]

We know what is dedupe and evaluated that we need that. We have a good success using NetApp dedupe and performance does not affect the servers even we are CIFS or iSCSI. We want to test out how ZFS perform. The fastest we can go is 5MB/s. The RAM usage is 37% of total 8GB and CPU is like only less than 5%.
That very minimum resources usage and yet ZFS perform very poorly.