Help cutting my FreeNAS teeth

[victorzamora]

Hey guys, so I believe I've gotten everything set up properly. Thanks to everybody for the guides, the help, and the posts. It's been incredible so far. I'm stuck at one point and am REALLY frustrated about what to do. I'm transferring data via Gigabit network from my main desktop to my FreeNAS box. I have the latest FreeNAS installed via USB on my FreeNAS machine. So, hardware:

Desktop:
FX-8320, Gigabyte GA-990FXA-UD3 mobo, AMD 7950, 2x4GB 1333MHz RAM, 1x80GB SSD, 1x1TB Samsung F3 (via USB cradle for now), 1x2TB WD Green (via SATA), Windows 7 Enterprise.

FreeNAS Box:
FX-4100, Gigabyte 78LMT-USB3 Mobo, onboard graphics, 2x4GB 1333MHz RAM, 4x2TB WD green (all via SATA) in a RAIDZ setup. OS installed on USB2.0 flash drive.

Router: Netgear WNDR3700. It's rated Gigabit, and both of my mobos are rated Gigabit.

So, I believe that one of my WD 2TB Greens in the NASBox might be going bad. I will replace it with the WD 2TB that's currently in my desktop. All of my data is in the one on my main desktop....so I'm going to switch it out later. My plan is to make the possibly bad one a separate drive on the NASBox for backing up my laptop, my fiancee's laptop, and my desktop. If it goes out, I'll replace it. I'll also be replacing it (or mirroring it with a new one) soon(ish) even if it stays good.

I forgot to mention, I'm transferring at ~6MB/s right now. That just seems crazy slow to me.

 

[gpsguy]

Given what you've said, I'd stop the data transfer now.

If you suspect a bad drive in your NASBox, then you need to find the problem and fix it before you copy/move any more data to it. And, if you do discover a bad drive, either RMA it or take the platters out and make a windchime for your fiancée.;)

Do a "zpool status -v" and post the results in code tags. We may suggest other tests too.

After we confirm your FreeNAS array is okay, we can look at other stuff. For example, if you have a Realtek NIC, the hamsters might be tired of pushing data around.

 

[victorzamora]

I ran "Zpool status -v" in the shell via the browser-based GUI and got a whole bunch of lines. The important one: No known data errors. It also said "No scan requested" and I got zeros everywhere. I'll include a screenshot in a second.
Also, thanks for the fast response!

Edit: Including picture

 

[victorzamora]

My desktop mobo has a Realtek 8111E LAN chip. My NASBox mobo has an unspecified Realtek GbE LAN chip.

 

[gpsguy]

At a minimum, I'd get an Intel Pro/1000 CT for your NASBox. An OEM version of it costs about $35, here in the US.

My NASBox mobo has an unspecified Realtek GbE LAN chip.

 

[victorzamora]

I understand, and I very well might....but is there a reason? This link shows that the NIC in my desktop isn't really any worse than an Intel NIC. I'm starting to think that the NIC in my NASBox mobo is a Realtek GbE (and that's the model name) NIC chip.
http://forums.freenas.org/threads/intel-nic-vs-realtek-nic-performance-testing.10325/

So, even if I do get one....it doesn't help me short term. Is there anything else to check or do or look at?

 

[gpsguy]

Assuming your devices are ada0 - ada3, I'd run smartctl -a /dev/adax on each one of them. Substitute the drive number for each of them, for example ada0. This will produce lots of output, about each drive.

I'd suggest you enable SSH, download PuTTy for your Windows machine and learn how to use it. Enabling SSH is covered in the manual.

This combination will allow you to connect to your FreeNAS machine. And, allows you to capture screen output, easily to a file.

 

[gpsguy]

In general, the Realtek NIC's might work okay for Windows (which supports lots of hardware). FreeNAS device support is more limited. Search the forum for "hamster" or "Realtek" and get an idea of the problems user's have with them.

I understand, and I very well might....but is there a reason?

There are a number of other items that might affect your performance. You never mentioned which version of FreeNAS you're using. We need to know the version? And, given your hardware, are you using the 64 bit version? Some folks accidentally install the 32-bit version and find out it's not using all the RAM, etc.

 

[victorzamora]

Sorry, I'm running 8.3.1....and yes, I'm definitely running the 64-bit version as it recognizes my 8GB of RAM instead of the 3.7 it would be limited to with x86. The "Network" tab shows this as my build: FreeNAS-8.3.1-RELEASE-p2-x64 (r12686+b770da6_dirty)

I'll look up "hamster" and "Realtek" to check FreeNAS-specific issues, but the link I found was from Dec 2012 and was using FreeNAS for the comparisons.

 

[gpsguy]

Once we're ready to look at performance, you'll need to run some tests on your array and from your workstation to the NASBox.

 

[victorzamora]

So, I'm working on getting Putty working....but in the meantime I ran the command directly in the Shell via Web-GUI on my NASBox. (I forgot to mention I can connect the NASBox to a monitor and keyboard if that helps.)

Yes, my disks are ada0-ada3. I ran the commands you suggested above and got consistent results for ada1-ada3, and a longer/more detailed result when I ran it for ada0.

One last thing: I know this won't help with proving "acceptance" by FreeNAS....but I've done network transfers before with MUCH higher speeds than 5MB/s.

 

[gpsguy]

Yes, I remember that thread. In the OP's case, he knew which NIC he had. Yours in an "unknown" one.

... but the link I found was from Dec 2012 and was using FreeNAS for the comparisons.

 

[cyberjock]

If you don't know how to interpret the smartctl output post it and someone will review it for you...

 

[victorzamora]

I'm trying to figure out how to post it :D

Until then, the one line that's worrying me is:

Code:

 04 61 0c 00 00 00 a0 Device Fault; Error: ABRT Commands leading to the command that caused the error were: CR FR SC SN CL CH DH DC Powered_Up_Time Command/Feature_Name -- -- -- -- -- -- -- -- ---------------- -------------------- ef 03 0c 00 00 00 a0 00 00:25:44.089 SET FEATURES [Set transfer mode] b0 d8 01 01 4f c2 e0 00 00:25:26.143 SMART ENABLE OPERATIONS SMART Self-test log structure revision number 1 No self-tests have been logged. [To run self-tests, use: smartctl -t] SMART Selective self-test log data structure revision number 1 SPAN MIN_LBA MAX_LBA CURRENT_TEST_STATUS 1 0 0 Not_testing 2 0 0 Not_testing 3 0 0 Not_testing 4 0 0 Not_testing 5 0 0 Not_testing Selective self-test flags (0x0): After scanning selected spans, do NOT read-scan remainder of disk. If Selective self-test is pending on power-up, resume after 0 minute delay.

It's all I can read as whatever's above it I can't scroll to. I can read stuff below, but it doesn't **SEEM** important.

Edit- I've Added everything I can see when I run it for ada0

 

[cyberjock]

You can take the output of any command and forward it to a file like this...


smartctl -a /dev/ada0 > /mnt/tank/ada0smartdata.txt

 

[gpsguy]

As a test, copy a big file (several Gb's) from the SSD on your pc, to FreeNAS and see what your performance is like.

One last thing: I know this won't help with proving "acceptance" by FreeNAS....but I've done network transfers before with MUCH higher speeds than 5MB/s.

 

[victorzamora]

Awesome, thanks! I looked through the txt file and it means very little to me. It's formatted poorly, so let me know if I should do something to fix it. Either way, I'll upload the file. I'll edit this post and upload the file for ada1 as well, just for comparison.

 

[victorzamora]

So I transferred TO my SSD using a WD Green via SATA at: 270MB/s max, ~120MB/s average. Total time: ~second for a single 8.73GB file.

Transferring to my NAS from my SSD is: estimated at taking 1 hour and is transferring at 3.23MB/s. The speed dropped to 2.60MB/s in a very short amount of time.

Two notes, the "discovering" portion of the transfer took NOTICEABLY longer from my SSD to my NAS than from my HDD to my SSD. Also, the speed always seems to deteriorate. It'll start high and then slowly drop off if I'm transferring via LAN. However, the speed seems to have picked back up to ~6MB/s.

 

[gpsguy]

ada0 reports that it's "failing now" on this entry.

Code:

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x002f   034   034   051    Pre-fail  Always   FAILING_NOW 5645

 

[cyberjock]

ada0 looks questionable. It has 3 reallocated event count and 3 current pending sector count. Both of those are a good indicator of a bad drive. If the number isn't changing and the hard drive can complete a zpool scrub and long test then I would continue to trust the drive.

ada1 has 16 for multi_zone_error_rate. I've had that number be non-zero and not had problems. So I tend to think that ada1 is fine. Just like ada0, if a zpool scrub is fine and a long test performs successfully then I wouldn't worry about that drive.