My Cloud 4TB - /etc/samba/smb-global.conf

Can someone please post the contents of the factory default  /etc/samba/smb-global.conf file for a My Cloud 4TB drive?

Thanks!

1 Like

You can retrieve the original from the firmware update file sq-030401-230-20140415.deb.zip.

/etc/samba/smb-global.conf:

[global]
load printers = no
printable = no
log file = /var/log/samba/log.smbd
max log size = 50
deadtime = 30
enable core files = no
security = user
encrypt passwords = yes
passdb backend = smbpasswd:/etc/samba/smbpasswd
create mask = 0775
force create mode = 0775
directory mask = 0775
force directory mode = 0775
local master = yes
domain master = no
preferred master = auto
os level = 5
use sendfile = yes
dns proxy = no
idmap uid = 10000-65000
idmap gid = 10000-65000
admin users =
null passwords = yes
map to guest = bad user
guest account = nobody
force group = share
unix extensions = no
acl check permissions = false
max protocol = SMB2
browseable = yes
syslog = 1
socket options = TCP_NODELAY IPTOS_LOWDELAY SO_RCVBUF=2048000 SO_SNDBUF=2048000
min receivefile size = 16384
smb encrypt = disabled
writeable = yes
delete veto files = true

Note: be careful when setting “dos filemode=yes”. We have seen where Samba does not

handle this properly (and crashes) upon a rename of a file where the authenticated user is not the owner.

##dos filemode = yes

 

Nazar78 wrote:

You can retrieve the original from the firmware update file sq-030401-230-20140415.deb.zip.

 

/etc/samba/smb-global.conf:

[global]
load printers = no
printable = no
log file = /var/log/samba/log.smbd
max log size = 50
deadtime = 30
enable core files = no
security = user
encrypt passwords = yes
passdb backend = smbpasswd:/etc/samba/smbpasswd
create mask = 0775
force create mode = 0775
directory mask = 0775
force directory mode = 0775
local master = yes
domain master = no
preferred master = auto
os level = 5
use sendfile = yes
dns proxy = no
idmap uid = 10000-65000
idmap gid = 10000-65000
admin users =
null passwords = yes
map to guest = bad user
guest account = nobody
force group = share
unix extensions = no
acl check permissions = false
max protocol = SMB2
browseable = yes
syslog = 1
socket options = TCP_NODELAY IPTOS_LOWDELAY SO_RCVBUF=2048000 SO_SNDBUF=2048000
min receivefile size = 16384
smb encrypt = disabled
writeable = yes
delete veto files = true

Note: be careful when setting “dos filemode=yes”. We have seen where Samba does not

handle this properly (and crashes) upon a rename of a file where the authenticated user is not the owner.

##dos filemode = yes

 

 

Those buffer sizes are hilarious for a system with so little RAM available.

Or engineers at WD are genius??

Eli_Tsigantes wrote: 

 

Those buffer sizes are hilarious for a system with so little RAM available.

Or engineers at WD are genius?? 

“Hilarious?”   Yeah, it’s funny that I can get 110 Megabytes/second performance on this NAS, so obviously those buffer settings are WAY too large.  :laughing:

so what would you recommend the buffer sizes be set to?

JohnQuintas wrote:

so what would you recommend the buffer sizes be set to?

I know you’re asking “Eli,” but my recommendation is to leave them where they are.

If you want to “tweak” them, despite what “Eli” says, the maximal performance change with those parameters is more dependent on your network environment than the memory on the Cloud.   

He doesn’t understand that it’s not consuming memory in the way he thinks, it’s a TCP SOCKET option that controls how much data Samba passes back and forth to the TCP socket.

TonyPh12345 wrote:


Eli_Tsigantes wrote: 

 

Those buffer sizes are hilarious for a system with so little RAM available.

Or engineers at WD are genius?? 


“Hilarious?”   Yeah, it’s funny that I can get 110 Megabytes/second performance on this NAS, so obviously those buffer settings are WAY too large.  :laughing:

Wow bro 110MB/s w/o jumbo frames is quite impossible to achieve as this device doesn’t support`em. My peaks are 80MB/s ups and 90MB/s downs. Stable downs but fluctuating ups even though these’s less to none retransmissions. Mind sharing your secret recipe? :flushed:

Nazar78 wrote:  Wow bro 110MB/s w/o jumbo frames is quite impossible to achieve as this device doesn’t support`em. My peaks are 80MB/s ups and 90MB/s downs. Stable downs but fluctuating ups even though these’s less to none retransmissions. Mind sharing your secret recipe? :flushed:

The fastest PC on my network happens to be a cheap little Windows 8 box I bought for my kids last year.   It can do 110MB/sec reads (for large files, of course) quite easily.

Win8.png

Both my Windows 7 desktop and my Linux workstation average around 80-90 MB/sec reads.

I’m guessing my network is a little out of the norm in that ALL of my switches support 802.3x Flow Control (as does the Cloud NAS), which might help somewhat…   Looking at the statistics on my switches, there are 802.3x Pause Frames being used, so it’s doing something…

I dont know man. I can never get more than 85MB/sec. The picture shows a file copy from the my cloud drive to a SSD drive capable of 500MB writes, running on a high end windows 7 PC. I guess i can blame my 70ft lan cable for some degration in speed or my router?

 

alirz1 wrote:

 I guess i can blame my 70ft lan cable for some degration in speed or my router? 

Only if it’s causing errors…   Length of cabling is irrelevant otherwise…

@TonyPh12345 You got this w/o tweaking the NAS? Cool! My dlink switch are flow control capable too. Guess I’ll have to look  further into my Win8.1 box. I’m concerned particularly about the fluctuating 30-80MB/s cat6 up stream to the NAS. May be obsolete tweaks which needs a tcp stack reset…

Thanks for sharing man!

  

 TonyPh12345 wrote:

Eli_Tsigantes wrote: 

 

Those buffer sizes are hilarious for a system with so little RAM available.

Or engineers at WD are genius?? 

“Hilarious?”   Yeah, it’s funny that I can get 110 Megabytes/second performance on this NAS, so obviously those buffer settings are WAY too large.  :laughing:


I know from reading old postings that you are one of the handful lucky people whose MyCloud miraculosely runs like a charm. And by repeating this you became a Honored Contributor.

No offense intended, but why should I believe a screenshot from you if the number of buyers complaining speed like me goes into the hundreds?

You lack the success story proving that your fantastic knowhow helped a dozen users to get the same fantastic speed you publish.

Regards from Greece

Eli

TonyPh12345 wrote:


JohnQuintas wrote:

so what would you recommend the buffer sizes be set to?


I know you’re asking “Eli,” but my recommendation is to leave them where they are.

 

If you want to “tweak” them, despite what “Eli” says, the maximal performance change with those parameters is more dependent on your network environment than the memory on the Cloud.   

 

He doesn’t understand that it’s not consuming memory in the way he thinks, it’s a TCP SOCKET option that controls how much data Samba passes back and forth to the TCP socket.

 

Even if you are the biggest expert here you never can guess how and what I imagine.

My educated guess: Samba might ignore these values or limit them to a reasonable size.

Instead of harassing me you might better study the source of Samba :open_mouth:

  

It still perplexes me that why for some the wd cloud works fine(including me) and for others it has been a nightmare…

Not to mention my primary wd cloud drive (i have two) is pretty moded and running extra stuff and it still works great. No sleep issue etc… Despite wd2go web access not working.

Other thing to keep in mind is that people who dont have a problem with the drive, will most likely never end on this forum…

Eli, So any reccomendations on SO_RCVBUF SO_SNDBUF sizes? for a 4TB My Cloud device?

@JohnQuintas Do you have any speed issues?

I’m only getting 3MB/s - 5MB/s

Are you on a wired gigabit connection or wireless?

wired gigabit

PC?  Mac?