Hi, My speed is at the max about 30 Mb/s
The switch is completely plain configured.
Only this
Set IP adres 172.16.0.1/24
Ether1 - Ether23 set to master ether24
That's it.
I connect a Supermicro server on port ether 2 and a Qnap-TS259 on ether9
Why is this slow?
Do I use fastpath with this config.

There is no fastpath in switch mode. Having all ports set as slave to one common port is correct setting. Maybe there is cable or autonegotiation problem...

Thanks,

I know for sure there is no cable problem. Alle new cables (30 centimeters long).
The same server and nas connected to a unmanaged gigabit switch gives me good speed but CRS125 not.

Post your export. Looks like some misconfuguration. Also provide the profile values when switch is under constant load reaching its limits.

This is the export.

# mar/02/2015 07:14:41 by RouterOS 6.27
# software id = E3Q0-ENNN
#
/interface wireless
set [ find default-name=wlan1 ] l2mtu=2290
/interface ethernet
set [ find default-name=ether1 ] master-port=ether24
set [ find default-name=ether2 ] full-duplex=no master-port=ether24 speed=\
1Gbps
set [ find default-name=ether3 ] master-port=ether24 speed=1Gbps
set [ find default-name=ether4 ] master-port=ether24
set [ find default-name=ether5 ] master-port=ether24
set [ find default-name=ether6 ] master-port=ether24
set [ find default-name=ether7 ] master-port=ether24
set [ find default-name=ether8 ] master-port=ether24
set [ find default-name=ether9 ] master-port=ether24 speed=1Gbps
set [ find default-name=ether10 ] master-port=ether24
set [ find default-name=ether11 ] master-port=ether24
set [ find default-name=ether12 ] master-port=ether24
set [ find default-name=ether13 ] master-port=ether24
set [ find default-name=ether14 ] master-port=ether24
set [ find default-name=ether15 ] master-port=ether24
set [ find default-name=ether16 ] master-port=ether24
set [ find default-name=ether17 ] master-port=ether24
set [ find default-name=ether18 ] master-port=ether24
set [ find default-name=ether19 ] master-port=ether24
set [ find default-name=ether21 ] master-port=ether24
set [ find default-name=ether22 ] master-port=ether24
set [ find default-name=ether23 ] master-port=ether24
/ip pool
add name=TEST ranges=172.16.0.2-172.16.0.254
/ip dhcp-server
add address-pool=TEST disabled=no interface=ether24 name=TEST
/port
set 0 name=serial0
/tool user-manager customer
set admin access=\
own-routers,own-users,own-profiles,own-limits,config-payment-gw
/ip address
add address=192.9.201.139/24 interface=ether20 network=192.9.201.0
add address=172.16.0.1/24 interface=ether24 network=172.16.0.0
/ip dhcp-server network
add address=172.16.0.0/32 dns-server=192.9.201.5 gateway=172.16.0.1 netmask=\
24
/ip route
add distance=1 gateway=192.9.201.250
/system clock
set time-zone-name=Europe/Amsterdam
/tool user-manager database
set db-path=user-manager

You have port 2 (one of your endpoints) set to half-duplex. That will slow things down as it prevents the device from sending and receiving simultaneously. You've got ports 2 and 9 hardcoded at 1 Gbps. While that shouldn't cause issues, it's usually easier and more beneficial to allow all devices to auto-negotiate link speed. Assuming that both devices are running at the 1 Gbps link speed, you should have wire-speed communication (brought down a bit by the half-duplex on port 2). If you're not getting full speed, you might want to look into other limitations, such as the hardware capabilities on each endpoint. It sounds to me like you're trying to transfer data from a server to a NAS. Once you take TCP overhead out of the equation, wire-speed 1 Gbps results in a max of about 100 MB/s. What's your hard drive setup look like on both the server and the NAS?

I just researched the Qnap...I assume you're running the 2 drives as a RAID 1 (mirror) for data protection? Your read/write speed is essentially the same as a single drive, so you're probably not going to get much more than 60-80 MB/s through to the hard drives themselves. Qnap quotes up to about 115 MB/s in testing, but that's probably as a RAID 0 array where you can write to both drives, plus any number of undisclosed optimizations.

Hi,

Thanks for your reply.
I see that ether2 is set to half-duplex. But I did not do that. So that is strange.
But I can not change it also.
Is this maybe the problem?
On the server the network interface is set to auto-negotiate and full duplex.
But the CRS125 set the port to halfduplex automaticly.
Weird.

Hi,

Thanks for your reply.
I see that ether2 is set to half-duplex. But I did not do that. So that is strange.
But I can not change it also.
Is this maybe the problem?
On the server the network interface is set to auto-negotiate and full duplex.
But the CRS125 set the port to halfduplex automaticly.
Weird.

the half-duplex on ether2 is no problem if autonegotiation is ON (and not configured to offer half duplex modes only) and the real link status showd you are using 1bgps/full.
If you want to change the settings, disable autonegotiation for the port temporarily, change the dulex settings and the allow the autonegotation again.

I have just tested CRS125 with very similar setup (ether1 used as a master port) and I had no problem to reach 980mbps of UDP packets throught it.

I think you should check ethernet port error counters on related ports on CRS and on the attached devices. Maybe there are (receive) errors (FCS etc)

Hi

This are the counters:

[admin@SW01] > /interface ethernet print stats-detail from=ether2
Flags: X - disabled, R - running, S - slave
0 RS name="ether2" driver-rx-byte=0 driver-rx-packet=0 driver-tx-byte=0 driver-tx-packet=0 rx-bytes=69 899 596 rx-too-short=0 rx-64=261 rx-65-127=298 rx-128-255=463
rx-256-511=409 rx-512-1023=257 rx-1024-1518=45 776 rx-1519-max=0 rx-too-long=0 rx-broadcast=69 rx-pause=0 rx-multicast=245 rx-fcs-error=0 rx-align-error=0
rx-fragment=0 rx-overflow=0 tx-bytes=1 811 190 tx-64=17 429 tx-65-127=7 769 tx-128-255=456 tx-256-511=1 tx-512-1023=8 tx-1024-1518=0 tx-1519-max=0 tx-too-long=0
tx-broadcast=66 tx-pause=0 tx-multicast=43 tx-underrun=0 tx-collision=0 tx-excessive-collision=0 tx-multiple-collision=0 tx-single-collision=0
tx-excessive-deferred=0 tx-deferred=0 tx-late-collision=0 tx-queue0-packet=101 tx-queue0-byte=6 511 tx-queue1-packet=25 562 tx-queue1-byte=1 804 679
tx-queue2-packet=0 tx-queue2-byte=0 tx-queue3-packet=0 tx-queue3-byte=0 tx-queue4-packet=0 tx-queue4-byte=0 tx-queue5-packet=0 tx-queue5-byte=0 tx-queue6-packet=0
tx-queue6-byte=0 tx-queue7-packet=0 tx-queue7-byte=0 tx-all-queue-drop-packet=0 tx-all-queue-drop-byte=0 tx-queue-custom0-drop-packet=0 tx-queue-custom0-drop-byte=0
tx-queue-custom1-drop-packet=0 tx-queue-custom1-drop-byte=0 policy-drop-packet=0 custom-drop-packet=0 current-learned-count=0 learn-failed-count=0
[admin@SW01] >