Tried JUMBO Frames still getting slow LAN

[philmiami]

I changed my wiring and put my QNAP TS-419P+ (with latest firmware)
EVERYTHING has jumbo frames enabled
my iMac ethernet
the LAN 3&4 on the RT-AC68U router
the QNAP Ethernet 1+2 (in fail over mode if 1 goes out but only uses eth1 plugged into LAN 3&4 on router)
A dlink green 8port dumb 1Gb switch that the iMac is plugged into and then plugs into LAN1 on router

The fastest I get is like 33MB/s
This is no were near 1Gb speeds
1GigaBIT = 128 MegaBYTE
so I am only getting around 23-25% of the speed
everything is set for 1000Mb FULL duplex (1000Mb TX/1000Mb RX)

Is there a way to check the LAN on the router is operating at 1000 Mb FULL duplex?

I wonder if the stupid switch is holding it back? Doubt it, as I think the overhead would at most be 5-7% running full duplex because it does not route, it multiport broadcasts.

The only routing is the router itself. How could I check?
The traffic manager screen is just for going out to eth0/wan
and doesn't show internal routing.
I am using the NAS as a Time Machine backup.
I mean, I am just wondering. This is not mission critical BUT to increase internal lan speed would make the process take less time.
The Qnap has Seagate Baracuda 7200RPM 2TB drives x 4 in a RAID 10 config. So I know my drives are fast enough.

 

[KGB7]

In the router go to Tools and scroll down it will show connection speed.
Also turn QOS off, HW acceleration On.


Also bypass the switch and connect directly to router.


What type of cabling are you using?

 

[thiggins]

The TS-419P+ has a Marvell 6282 processor. That class of NAS is not going to reach full Gigabit speed. They'll top out at more like mid 50 MB/s for Windows file copy.

BUT you must be using a machine for the test that can support that rate of throughput. So first look at the throughput that the machine that you are using for the test can achieve.

And disable Jumbo frames. They aren't really going to help you with today's bus architectures and adapters.

 

[drabisan]

Are you measuring the speed achieved by Time Backup? If so, that's wrong in so many levels and you should be happy with the figure! Time Backup is "copying" lots of small files into also files on your storage.
Other think you should consider: OS version. Apple chose to optimize SMB lately. But TB is using AFP. And AFP was never the best for any non-Apple NAS'.

For what it matters, 68U can easily switch 1G between 2 LAN ports. I test this for my pleasure and it was easy to reach 1G. Very large files, SMB and powerful machines. I didn't need jumbo frames. But tests were made with i7 CPU with SSD and lots of RAM! I can reach 1G between my 2 Synology boxes only with jumbo frame, but combined both Synologys plus HDDs on it are not that expensive/powerful as either machines from the first test.

 

[Preskitt.man]

I perform basic LAN performance tests often at work for some of networked environment. A single high end Dell servers over GE going our high end NAS device will rarely achieve GE speeds (128 MBs).
* Network overhead itself will remove up to 10%
* Many file copy utilities are single threaded. You will need a multi-threaded file copy utility to maximize performance over the GE link.
* When testing, obviously important to make sure everything else on both the client and server are quiesced.
* Cycles (performance) will also be taken up as the client reads data from disk and client writes data to disk. If you truly want test raw synthetic performance, read from /dev/null and write to the bit bucket.
* Consumer grade appliances are rarely designed for maximum network performance.

 

[philmiami]

1: Yes thru Time Machine but I excluded all directories except /user
2: I know there is overhead in networking yes, but what I am looking at is coming from my iMac in HSF going over the Gb internal to my QNAP in EXT4 file format.
I could say that would be another 20% loss, so taken say 35% loss, should still equal out to around 50-60 MB/s
not coming out at the 30-33 MEGABytes/s
so there is more loss there than what I figured would be
I am going to probably switch the cables out too.....
ah, well as long as it backs up

 

[5mall5nail5]

I changed my wiring and put my QNAP TS-419P+ (with latest firmware)
EVERYTHING has jumbo frames enabled
my iMac ethernet
the LAN 3&4 on the RT-AC68U router
the QNAP Ethernet 1+2 (in fail over mode if 1 goes out but only uses eth1 plugged into LAN 3&4 on router)
A dlink green 8port dumb 1Gb switch that the iMac is plugged into and then plugs into LAN1 on router

The fastest I get is like 33MB/s
This is no were near 1Gb speeds
1GigaBIT = 128 MegaBYTE
so I am only getting around 23-25% of the speed
everything is set for 1000Mb FULL duplex (1000Mb TX/1000Mb RX)

Is there a way to check the LAN on the router is operating at 1000 Mb FULL duplex?

I wonder if the stupid switch is holding it back? Doubt it, as I think the overhead would at most be 5-7% running full duplex because it does not route, it multiport broadcasts.

The only routing is the router itself. How could I check?
The traffic manager screen is just for going out to eth0/wan
and doesn't show internal routing.
I am using the NAS as a Time Machine backup.
I mean, I am just wondering. This is not mission critical BUT to increase internal lan speed would make the process take less time.
The Qnap has Seagate Baracuda 7200RPM 2TB drives x 4 in a RAID 10 config. So I know my drives are fast enough.

ping -l -f 7000 [IP of QNAP]

 

[philmiami]

ping -l -f 7000 [IP of QNAP]

iMac
-f is invalid argument
_____________________
phil-lan-imac:~ philmiami$ ping
usage: ping [-AaDdfnoQqRrv] [-b boundif] [-c count] [-G sweepmaxsize] [-g sweepminsize]
[-h sweepincrsize] [-i wait] [-l preload] [-M mask | time] [-m ttl]
[-p pattern] [-S src_addr] [-s packetsize] [-t timeout]
[-W waittime] [-z tos] host
ping [-AaDdfLnoQqRrv] [-c count] [-I iface] [-i wait] [-l preload]
[-M mask | time] [-m ttl] [-p pattern] [-S src_addr]
[-s packetsize] [-T ttl] [-t timeout] [-W waittime]
[-z tos] mcast-group
phil-lan-imac:~ philmiami$ ping -l 7000 192.168.0.29
ping: -l flag: Operation not permitted
phil-lan-imac:~ philmiami$

 

[5mall5nail5]

iMac
-f is invalid argument
_____________________
phil-lan-imac:~ philmiami$ ping
usage: ping [-AaDdfnoQqRrv] [-b boundif] [-c count] [-G sweepmaxsize] [-g sweepminsize]
[-h sweepincrsize] [-i wait] [-l preload] [-M mask | time] [-m ttl]
[-p pattern] [-S src_addr] [-s packetsize] [-t timeout]
[-W waittime] [-z tos] host
ping [-AaDdfLnoQqRrv] [-c count] [-I iface] [-i wait] [-l preload]
[-M mask | time] [-m ttl] [-p pattern] [-S src_addr]
[-s packetsize] [-T ttl] [-t timeout] [-W waittime]
[-z tos] mcast-group
phil-lan-imac:~ philmiami$ ping -l 7000 192.168.0.29
ping: -l flag: Operation not permitted
phil-lan-imac:~ philmiami$

-s instead

 

[philmiami]

phil-lan-imac:/ philmiami$
phil-lan-imac:/ philmiami$
phil-lan-imac:/ philmiami$ ping -s7000 192.168.0.29
PING 192.168.0.29 (192.168.0.29): 7000 data bytes
7008 bytes from 192.168.0.29: icmp_seq=0 ttl=64 time=0.632 ms
7008 bytes from 192.168.0.29: icmp_seq=1 ttl=64 time=0.686 ms
7008 bytes from 192.168.0.29: icmp_seq=2 ttl=64 time=0.709 ms
7008 bytes from 192.168.0.29: icmp_seq=3 ttl=64 time=0.679 ms
7008 bytes from 192.168.0.29: icmp_seq=4 ttl=64 time=0.713 ms
7008 bytes from 192.168.0.29: icmp_seq=5 ttl=64 time=0.678 ms
7008 bytes from 192.168.0.29: icmp_seq=6 ttl=64 time=0.698 ms
7008 bytes from 192.168.0.29: icmp_seq=7 ttl=64 time=0.726 ms
7008 bytes from 192.168.0.29: icmp_seq=8 ttl=64 time=0.684 ms
7008 bytes from 192.168.0.29: icmp_seq=9 ttl=64 time=0.628 ms
7008 bytes from 192.168.0.29: icmp_seq=10 ttl=64 time=0.676 ms
7008 bytes from 192.168.0.29: icmp_seq=11 ttl=64 time=0.710 ms
7008 bytes from 192.168.0.29: icmp_seq=12 ttl=64 time=0.666 ms
7008 bytes from 192.168.0.29: icmp_seq=13 ttl=64 time=0.718 ms
7008 bytes from 192.168.0.29: icmp_seq=14 ttl=64 time=0.720 ms
7008 bytes from 192.168.0.29: icmp_seq=15 ttl=64 time=0.709 ms
^C
--- 192.168.0.29 ping statistics ---
16 packets transmitted, 16 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 0.628/0.690/0.726/0.028 ms