Test done from a 1st gen MacBook (core duo) over gigabit ethernet with jumbo frames enabled on both the NAS and the Mac (and verified to actually work). The mac is plugged into an 8 port gigabit netgear switch which then runs to a 4 port gigabit netgear (WNDR3700v2) which the ReadyNAS is plugged into.
I'm pretty happy with those #s and I'd assume I'll get better write speeds as I add more disks
I have Teracopy installed. My read speeds are about 37-38MB/sec. Sometimes they spike as high as 47MB/sec, but it always bounces back to 37-38MB/sec right after.
Write speeds are much worse, but that could be due to how full the array is. Currently they average about 8-15MB/sec. I can remember them being as high as 25MB/sec under XP - but the array was almost empty when I first tested it.
No matter what I try I cannot get tranfer speeds better than 11MB/s. I have changed cables, even disabled options that were recommended to be disabled to improve speed. The network status on my NAS frontview page says it is "Online / 1000 Mbit / full duplex but still only getting 11Mb / second transfer speeds from my internal hard drives to the NAS.
I have to say that, unfortunately, the industry is VERY SLOPPY about performance/speed where networking is concerned.
If the syntax of your abbreviations is correct, then you are getting a good rate.
MB and Mb are two completely different things, but constantly used interchangeably. They're not, never were. Microsoft is terrible about this and their bad habits have become the worlds. Performace in networking is ALWAYS measured in megabits, except by some companies such as Microsoft, who measures it in megabytes. Cheifly, because microsoft has always been sloppy to bad at networking.
There is no accurate conversion between Mb & MB because the first is basic measure of transmission (layer 1) and the other is the result from the perspective of an application (layer 7). To transfer 1 MB of data does NOT take 8Mb. And different protocols and apps will make that ratio vary. But because there is no right answer, we typically use say 1MB is equal to 8Mb. It is a guesstimate that gets you somewhere in the ballpark, but always short of the real Mb number. How short--you cant know.
So the 11MB rate you mention would be more than 88Mb on the wire. Over 60Mb there are always other factors that are probably constraining transfers. Are both client and server running at 1000 or 100? Are they on the same switch? If you are measuring a transfer of files what is the write speeds of the destination as write is always slower than read. Competing tasks? etc.
Hi all, Just got a Duo V2 the other day and wanted to check in on the write speeds over a gigabit connection. I get between 57-60MB/s on a large MKV file, just wanted to check that sounds about right I couldnt find too many benchmarks online for it?
If you are able to get those write speeds then you are doing good. From the reviews I've read, most seem to top out at 35-50 MB/s write and ~90 MB/s read.
WRITE Total I/O per second = 2.56 Total MB per second = 0.64 Average I/O Response time (ms) = 358.4287 Maximum I/O response time = 2515.4065 % CPU Utilization (total) = 4.77% Total error count = 1
READ Total I/O per second = 2.64 Total MB per second = 0.66 Average I/O Response time (ms) = 378.0300 Maximum I/O response time = 6039.6648 % CPU Utilization (total) = 3.55% Total error count = 0
I have a NASDUO V2 with a (for the moment) single 2tB WD Caviar Green (WD20EARX) installed.
Very disappointing and I don't know what to do. Running the Iometer was a huge step in my knowledge.
OK thought I would update you on a change made to my NIC from '100Mbps / Full Duplex' to 'Auto Negotiate 1000mbps'. Revised results in brackets....
WRITE Total I/O per second = 2.56 [201.02] Total MB per second = 0.64 [50.26] Average I/O Response time (ms) = 358.4287 [4.79] Maximum I/O response time = 2515.4065 [143.53] % CPU Utilization (total) = 4.77% [44.13%] Total error count = 1 [0]
READ Total I/O per second = 2.64 [169.45] Total MB per second = 0.66 [42.36] Average I/O Response time (ms) = 378.0300 [5.89] Maximum I/O response time = 6039.6648 [83.72] % CPU Utilization (total) = 3.55% [30.21] Total error count = 0 [0]
Single Caviar Green 3TB on NAS : -40MB/s Read&Write with CIFS/SMB access -55-90MB/s Read and 55MB/s Write with ftp access
Current config (2 Seagate 3TB added) : -45MB/s Read&Write with CIFS/SMB access -60-95MB/s Read and 60MB/s Write with ftp access
With the current config, IOMeter gives me 50MB/s Read and 55MB/s Write.
I don't know if jumbo frames are worth the investment on my Ultra4 (I would need to buy at least a compatible NIC on my computer and a third GS105 switch, my router is said to support only 2k jumbo frames). FYI, iperf gives me a constant 110MB/s between my computer and my NAS.
Hello folks, I started with the old 600A and I've had an X6 after that. I just picked up the latest Pro 6 and am transferring all my data using rsync from the X6 to the Pro 6 using a gigabit switch (which is connected to a 100Mb router for readyfront access).
Both devices say they have gigabit connections. The best speed I am seeing is 3MB/s with jumbo frames enabled, 2.5MB/s with it disabled. Think the X6 nic is just not having it?
I am under the impression that having the two ReadyNASisis on the gigabit switch should leave them unaffected by the 100Mb router but do correct me if I am wrong about that.
...I am under the impression that having the two ReadyNASisis on the gigabit switch should leave them unaffected by the 100Mb router but do correct me if I am wrong about that...
You are not wrong about that.
I think your speed issues are related to the CPU requirements for rsync.
@gattung - RSYNC is a memory and CPU hog when doing large tasks. It's great for incrementals though.
mdgm has a great guide when migrating. I believe he has a link in his signature. NFS is much faster for the initial transfer then you can switch to RSYNC for incrementals.
Just to make sure, did you upgrade your network cables when adding the gigabit switch? Should be at least CAT5e or higher.
Single Caviar Green 3TB on NAS : -40MB/s Read&Write with CIFS/SMB access -55-90MB/s Read and 55MB/s Write with ftp access
Current config (2 Seagate 3TB added) : -45MB/s Read&Write with CIFS/SMB access -60-95MB/s Read and 60MB/s Write with ftp access
With the current config, IOMeter gives me 50MB/s Read and 55MB/s Write.
I don't know if jumbo frames are worth the investment on my Ultra4 (I would need to buy at least a compatible NIC on my computer and a third GS105 switch, my router is said to support only 2k jumbo frames). FYI, iperf gives me a constant 110MB/s between my computer and my NAS.
I just installed a 2GB Corsair RAM module, I now get : -60MB/s Read and 48MB/s Write with CIFS/SMB access -59MB/s Read and 57MB/s Write with IOMeter.
The link on how to set up IOMeter seems to be dead, so i'll post ATTO disk benchmark result.
ReadyNAS Duo v2, SMB/CIFS, x2 SAMSUNG HD642JJ in Raid 1, Cat5E cables, Netgear WNDR3700v2 router and Windows 7 machine with integrated Realtek Gigabit NIC.
EDIT. Same test just with single Samsung HD204UI hard drive:
ReadyNas Ultra 2+ Connected to Apple Airport Extreme Router with Gigabit Ethernet Laptop Connected to Rouer with WiFi 2.4GHZ 802.11N
Average read/write speed over wireless = 11mb/s
I'm not trying to be rude, but you're testing capabilities of your (Apple Airport EXtreme) router and not Netgear ReadyNAS, iPaul81. Even my 4 years old NAS was able to provide such speeds. Over the LAN you would get much better/adequate results.
Well, due to a failed standard transfer over SMB (some files wouldn't be copied), i was in a hurry and retested transfer over FTP : 105-108MB/s constant read.
I don't know where was the bottleneck (my e-sata drive goes up to 105-110MB/s, the gigabit LAN with the NAS was rated constant 108-110MB/s) but I didn't care 😄
I'm wondering, is CIFS that slow, or do i have a misconfiguration ?
Tonight, i was transfering files with ftp while sabnzbd was DLing at full speed (2MB/s), the speed was 100+ MB/s. And on the other hand, i can't seen to exceed 50-60MB/s with CIFS.
I am definitely seeing a very large difference in write speed between the cache and the disks. When I initiate a write I get 800+ Mbit/sec for about 700MB of data, and then a halt while the NAS writes to disk, then a very consistent pattern of around 100Mbit/sec for 200MB of data and then halt as the NAS writes it, repeat, until the data is all sent. If I send less than 700MB, then it is like lightning, but once the cache is full my writes are much slower.
Mine? I have 6x Seagate Constellation ST32000644NS drives. No problems in SMART listed. I have Bonnie installed but it was unreliable so I disabled it.
