NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Ready NAS 104 Crashes when changing Network setting (creating new bond) ...
I was following this thread describing my issue on RN 104 - any attempt of creating ethernet bond (even latest OS 6.10.3) leads to crash displaying something 'putname+14' on the display. The NAS hang...
I did extensive tests (minimal traffic on network) with various combinations and even with X-RAID, one SSD drive (via 1GB ethernet Cat6 cable) I am getting avg 50MB/s for write and 30MB/s for read operations. Factory default - 1SSD 500GB - SMB access.
NAS performance tester 1.7 http://www.808.dk/?nastester
Running warmup...
Running a 400MB file write on \\10.24.1.110\Videa 5 times...
Iteration 1: 50,14 MB/sec
Iteration 2: 50,75 MB/sec
Iteration 3: 52,70 MB/sec
Iteration 4: 47,27 MB/sec
Iteration 5: 51,61 MB/sec
-----------------------------
Average (W): 50,49 MB/sec
-----------------------------
Running a 400MB file read on \\10.24.1.110\Videa 5 times...
Iteration 1: 28,58 MB/sec
Iteration 2: 30,16 MB/sec
Iteration 3: 30,93 MB/sec
Iteration 4: 34,08 MB/sec
Iteration 5: 28,32 MB/sec
-----------------------------
Average (R): 30,41 MB/sec
-----------------------------
CPU temp of NAS (mode set to Cool) is somewhere around 59 - 65 degrees of Celsius. I got an impression if I am getting over 65 degrees - I am having worse performance measurements.
If I run local speed tests using DD I am getting:
root@nas-E5-36-84:/data/Videa# dd if=/dev/zero of=/data/Videa/samplefile bs=1M count=4096 oflag=direct
4096+0 records in
4096+0 records out
4294967296 bytes (4.3 GB, 4.0 GiB) copied, 24.07 s, 178 MB/s
root@nas-E5-36-84:/data/Videa# dd if=/data/Videa/samplefile of=/dev/null bs=1M count=4096 iflag=direct
4096+0 records in
4096+0 records out
4294967296 bytes (4.3 GB, 4.0 GiB) copied, 21.7766 s, 197 MB/s
root@nas-E5-36-84:/data/Videa#
root@nas-E5-36-84:/data/Videa# dd if=/dev/zero of=/data/Videa/testfile bs=1M count=4096 oflag=dsync
4096+0 records in
4096+0 records out
4294967296 bytes (4.3 GB, 4.0 GiB) copied, 144.067 s, 29.8 MB/s
root@nas-E5-36-84:/data/Videa# dd if=/data/Videa/testfile of=/dev/null bs=1M count=4096 iflag=dsync
4096+0 records in
4096+0 records out
4294967296 bytes (4.3 GB, 4.0 GiB) copied, 27.2464 s, 158 MB/s
So unless you see something else to improve I am thinking that limiting factor is my NETGEAR WNDR4300v2 router ind this case ... ???
BTW: Seems factory default solved the crash with creating teaming bond of eth0/1. I tested the speed transfer of all modes and was getting basically almost same performance (+-10%) with all teaming modes except Broadcasting which was really bad. Also the performance without teaming and with teaming was basically same.
calaba wrote:
Average (W): 50,49 MB/sec
Average (R): 30,41 MB/sec
So unless you see something else to improve I am thinking that limiting factor is my NETGEAR WNDR4300v2 router ind this case ... ???
The PC running the test is also using gigabit ethernet? Or is it using WiFi?
The router (and PC) wifi would certainly limit your throughput. But it shouldn't be limiting it when all the connections are ethernet.
If you have internet security software running on the PC, you might try disabling it (also disabling any antivirus protection). That could be affecting the speeds.
Yes, the PC was connected via 1GB Ethernet Cat 6 cable.
The impact of antivirus SW won't be that bad + local tests on USB3 on the very same SSD used in the NAS were giving transfer 150 - 300MB/s so I can almost for sure eliminate the security SW.
Oops- original reply was cross-posted (not relevant to you).
calaba wrote:
The impact of antivirus SW won't be that bad + local tests on USB3 on the very same SSD used in the NAS were giving transfer 150 - 300MB/s so I can almost for sure eliminate the security SW.
It's very unusual for read speeds to be slower than write speeds. AV software will scan on file access, and that can slow down read speeds over the network. So I suggest trying it, and not assuming that it's not a factor.
Good point - with the read perf. slower than write - it always puzzled me :-).
I am rebuilding the NAS - backing up the data now - once done I will test RAID-6 (with 4 3TB drives, switched back from Flex-RAID to XRAID after adding 4th drive) and RAID-10 (with 4 3TB drives).
While testing the read/write performance for RAID-6/10 I will quick test also the Factory default X-RAID - WITHOUT the anitivirus apps running.
To explain the reason for XRAID-6/10 test -> I am thinking if I cannot make it faster - then use it with better reliability instead. So sacrify the performance and get beeter sleep to be protected against URE issue with RAID-5. I didn't find any good discussion comparing X-RAID against RAID-6/10 so I said to myself that if I cannot achieve the highest throughput anyway I might at least improve fault tolerance. Seems RAID-6 should be best but for RN104 might be too much work for CPU ... thus fallback to RAID-10 ... also thinking that if I disable X-RAID and use Flex-RAID then the additional layer of flexibility provided by X-RAID, if sacrified (not used), might ease up the CPU stress ... are those lines of my thoughts okay (for 4x3TB drive on ARM based RN104 model) ... ?? or would you recommend different RAID configuration (assuming here that I cannot make it fast anyway -> so will improve at least the reliability).
Also speaking of reliability - isn't RAID-6 (or RAID-10) with some Shares on the Volume having bit-rot-protection enabled (thus whole Volume having checksums enabled) too much of overkill for my RN104 ??
Thanx for any insights.
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
