Questions & troubleshooting my first TrueNAS

[Ibrahim05]

1.One of my HDDs (WD Red Pro NAS 18TB) is correctly recognized by my bios but when looking through the TrueNAS GUI it displays as 0B (image 1). Do I ignore this and pool it anyways or is there a fix for this?
2. The command line is constantly outputting this error (image 2). I am not sure if it affects functionality but should care about this?
3. I have 18TB of storage and I also have a combined 3.5TB of external and internal storage. I was wondering if there is some implementation out there to have 3.5TB of the total 18TB auto-backed up? Does such a configuration exist even on any platform?
4. Is there some way to create some functionality with the NAS similar to the recycle bin in Windows? My previous Ubuntu NAS had an issue where if a file was deleted via File Explorer it was always deleted forever. Nothing went wrong because of it (thankfully), but I'd like some extra protection for deleted files.
5. My connection to the TrueNAS GUI always says it's not secure. I assumed it wasn't important but is it?

 

[Arwen]

Here is what I can assist with;

1. How are the disks connected?
   Especially how is the "missing" disk connected?
   Is the "missing" disk used or was temporarily used elsewhere, (like in another RAID scheme)?
2. Do you have USB attached storage devices?
   TrueNAS & ZFS don't really support USB attached storage for data disks... Sometimes USB attached storage can be used for boot drives successfully, though don't use cheap USB flash drives.
3. I wrote a rough How To for backing up to locally attached disks. It's not perfect and it is not a meant to be perfect for new users. Just a guide of HOW it can be done, (and does work for me!).
   

   How to: Backup to local disks

   How to: Backups to local disks Overview: This subject comes up often enough that I thought I would start a post of how I do it. Specifically I back up my TrueNAS to a hot-inserted, locally attached disk, then when done, hot-remove and take...

   www.truenas.com
4. Don't know
5. If you access your TrueNAS from a local and secure network, it's fine. That is because the default protocol is HTTP. You would have to add your own SSL certificates to get HTTPS, which some people think is a pain in the rear.

Since you seem to be a new user, (1 forum post and just joined today), I'd suggest reading up on ZFS, suggested hardware and TrueNAS. See the Resource section, link at the top of any forum page, for a listing. Now some, well most, won't apply to new users, their are lots that do help new users. And even testing out TrueNAS in a VM like under VirtualBox.

 

[danb35]

You would have to add your own SSL certificates to get HTTPS, which some people think is a pain in the rear.

Well, TrueNAS includes a self-signed TLS certificate by default, which will let you use HTTPS, but you'll get a certificate error. To avoid that, as you say, you'd need to install your own cert, which does take some doing.

 

[Ibrahim05]

1. The two HDDs and the SSD (boot pool) are all just connected straight to the motherboard's sata ports and properly powered. The missing disk was fresh from the box for the most part. I did accidentally make all my drives as boot pools during the initial set up so that is the only thing I've explicitly done to the drive before removing it from the boot pool via the "boot pool status" menu.
2. No, I'm only currently working with internal drives. I tried seeing if I could add external HDDs to the pool and they properly showed in the menu, but I never did anything beyond check it showed up as an option.
3. Thanks for the resource! I'll mess around with backing up once I get the main share working.
5. Ok, as long as I'm not at risk I don't mind what my browser said so I'll leave that alone.

 

[joeschmuck]

Please follow the forum rules which is linked at the top of every page, and post your system information.

Answer for 4) Yes you can, It is called Snapshots but you cannot just restore the file as easy as a Windoze system would, it requires a bit more effort but you can recover files. Read the TrueNAS User Guide, it has a lot of information on this topic and many other things. Read it a second time and you will likely understand it more the second time around. I have read it quite a bit because I don't have a photographic memory. Geez, I wish I did at times.

Questions:
1) Has this NAS ever worked or is this a new build for you? If this is a new build, did you read the Hardware Guide?
2) The hard drives, are they all new and unused?
3) What is your boot drive?

Forum Rules

Welcome to the TrueNAS Community! Updated 11/28/2023 First and foremost I want to thank you for joining the TrueNAS community! My name is JoshDW19 and I'm the Community Manager here at TrueNAS. The TrueNAS Community is a friendly place to discuss open source software, technology, and products...

www.truenas.com

Hardware Recommendations Guide

This is the latest edition of the FreeNAS Community hardware recommendations guide. The current major version is R2, dated January 2021, with the last minor update on 2021-01-24. The format has moved away from the forum post form factor, to...

www.truenas.com

 

[Ibrahim05]

My mistake, here are the specs:
Motherboard- MS-7778 ver 1.0
CPU- AMD A8-5500 APU
13.4 GiB RAM
Boot Drive- Inland 256GB Professional
Drive- WD Red Pro NAS 18TB
Drive- 750 GB HDD (I am not sure what model)
Network Card- I am just plugged into the onboard Ethernet port

Thanks for informing me about the snapshots feature, I'll take a look at the discussions and guides that exist and get that working!

1) The system worked on Ubuntu but I never installed the 18TB to it, it's my first build with TrueNAS. I gave the hardware guide a quick look beforehand which helped with my initial confusions, but nothing specifically related to my issues that I could find.
2) The SSD boot drive and the 750 GB HDD have been in use for 1-2 years, but the 18TB HDD has not been used previously.
3) The SSD.



I did some messing around and apparently the server was still booting from the 18TB HDD, so I fixed that so it boots properly. There still is the strange "da3" disk
which appears under my list of disks even though there are only 3 disks attached (image 1). Problem 1 seems to be solved and I can just pool the 16.37TiB drive now and all the other drives are showing, should I just ignore the 0B drive? Apart from that my only remaining question is should I fix my problem from #2 of the initial post?

 

[joeschmuck]

The first steps you should do are:
1) Run Memtest86 or Memtest86+ (use the free version) and let it test your RAM for 3 to 4 complete passes, no errors are required.
2) Run a CPU stress test such as Mprime or Prime95 for 30 minutes minimum to get a feeling that the CPU is good. Run it overnight if you really want to test it, make sure your fans are running, the CPU will get hot.

Once those are good, and they must be good or you have to fix the problem before continuing, an unstable system will cause you problems.

If you have your email setup, it will make obtaining data much easier for you. Setup the root email account, make sure you can receive a test email. This is something new I'm doing to help people collect data easily.

3) If you do not have the email setup: in the GUI -> Shell -> type the command below and post the output.

Code:

smartctl -x /dev/ada0

3a) If you do have email setup: in the GUI -> Shell -> type the command below and an email with the text data will show up. Change the email address of course, that one is really not mine.

Code:

smartctl -x /dev/ada0 | mail -s “Drive ada0 SMART Data” joeschmuck@gmail.com

4) Run the command again but change "ada0" to "ada1", "ada2", and "da3" to have four different outputs. This will provide us more usable data on the drives in your system.

I suspect you plan to run stripe pools which you can force (checkbox) to happen, it of course it not how we like to see things done but it will work and you would have no redundancy.

Hopefully you read the entire message because I'm going to tell you, you can perform steps 3 and 4 now and post the results and then roll up to step 1 and do the stability testing. And yes, RAM erros do happen, I was doing some MemTest86 V10.6 Free on a system I plan to sell, and it failed. I was very surprised because it had passed the test last year. I will test my RAM periodically because it's an easy test and is a bit of a safe measure. I pulled out two sticks of RAM, all is passing, now I need to swap the sticks out for the other two and test those. If all tests fine, then all four go back in and I hope I just needed to reseat the modules, otherwise I likely have a memory controller problem, that means a bad motherboard.

It happens to the best of us.

Good luck and we can figure out what "da3" is. I would not ignore it.

 

[Arwen]

Some keyboards, USB hubs and even chassis include a memory card reader. Even if it is inaccessible, (my USB keyboard / mouse switch insists it has Gigabit Ethernet, except no port for it's use...).

So I would not be too concerned about "da3". This board does seem to have this, which might be it;

• One mini card

HP and Compaq Desktop PCs - Motherboard Specifications, (JasmineR) | HP® Customer Support

Specifications for the Jasmine-R motherboard

support.hp.com

 

[Ibrahim05]

So I would not be too concerned about "da3". This board does seem to have this, which might be it;

• One mini card

Yes! That was what da3 seemed to be. I removed the front io and da3 is not on the drive list anymore.

The first steps you should do are:

Now that I've found what da3 was, should I still do all the steps or which steps should I skip? I am hesitant to do your step 2 (CPU stress test) because I cheaped out on the cooling because I'm expecting the server to not get hot by doing exclusively file transfers. Do you think I should do it anyways?

 

[joeschmuck]

Run the MemTest86 on your RAM first. See if that is stable. If it is stable then run the CPU stress test for 5 minutes, see if the CPU heatsink get too hot to touch or use an infrared thermometer to read the CPU temp. The CPU actual temp will be several degrees higher. Your CPU max operating temp is 71C. Try not letting it get to 65C. TrueNAS will heat up the CPU during a Scrub, it's not much but if you have poor cooling, the CPU could reach it's high temp. It's better to know sooner than later.

If you are not storing important data on this NAS, you don't need to do any testing. However if there is a problem and you ask for help, we will need to know if the system passed the stability testing and we may ask you to run it again to verify your system is stable. Once that is done then we can rule out system stability as a problem.

But lets recap... Are all the immediate problems resolved? No more errors. If not, run those stability tests and then report those results, along with what problems you still have.

 

[Ibrahim05]

I ran MemTest86 and I got no errors except on the last test (images 1-3). I didn't communicate properly that by "cheaped out on cooling" I just meant I am using 1 case fan and a stock CPU cooler, so I don't have an exposed CPU heatsink to feel for temperature on. To perform the test and protect my CPU from overheating, what would I do (I don't have access to an infrared thermometer)?

Questions remaining:
1. Is the error "DMA write error" important or should I ignore it (if so how should I fix)?
2. How do I perform the suggested tests?

 

[joeschmuck]

1. Is the error "DMA write error" important or should I ignore it (if so how should I fix)?

DMA stands for Direct Memory Access. Test 14 is listed as experimental and since only that test failed, you might be okay BUT, I would now run the MemTest86+ test just to cover your bases. Here is the URL to grab a copy, be aware there are a lot of adds that say "Download" or what-not. Download the Latest Version and run this. Make sure it passes at least 4 complete times, or run it overnight. The goal is to ensure it does not fail. If it asks you to buy the program, stop and do not.

If MemTest86+ shows a failure, then unplug the system and remove the RAM and reinstall it. Test again.

I didn't communicate properly that by "cheaped out on cooling" I just meant I am using 1 case fan and a stock CPU cooler

Then you do not need to worry about the CPU heat if you have the stock cooler in place. Run the CPU stress test for 30 minutes, if it passes then you have a stable system, at least what we can test. If it fails then you may have a bad thermal conduction issue between the CPU and the heatsink, or your CPU is faulty. 30 minutes is reasonable in my book but as I said before, some people will run it for day, up to a month for a critical system.

If you can get both tests to pass, your system is considered stable, mostly. These tests basically are proofing the CPU/RAM/Power Supply/Motherboard. The other issue we have seen but it is more of a communications issue is if using a RealTek NIC. I didn't look at your motherboard to see if it has that or an Intel NIC. If you have RealTek, use it until you notice a data transfer issue, at that point you may need to install an Intel NIC add-on card and disable the onboard NIC.

Hopefully this helps move you forward.

 

[Ibrahim05]

Ok, so I ran MemTest86+ and verified that 2 passes had 0 errors. I tried doing 4 passes but the 2 passes already took ~24+ hours. When I checked on the progress at ~30hrs the power went out and the progress was lost. I assumed 2 passes over such a long time is probably good enough but if not I can run it again. Images 1 and 2 are of the info on the MemTest86+ screen.

I ran a CPU stress test on Ubuntu with the apps "stress" and "s-tui" and had no problems over 30 min, the graphs were essentially perfectly flat at 100% and the temps were low (I presume good cooling, open case, and low ambient temperatures help). Image 3 is not of the results at 30min (I forgot to take a picture before exiting) but the graphs did not change at all the entire time.

I am not sure about anything regarding NIC and I haven't checked any of that yet. I've never heard of a RealTek or Intel NIC so I am not sure how to check if I have one or which I have.

After now setting up the pools and data sharing I did realize I encountered some issues (may be related to the "DMA Write Error" from earlier):
1. I am having issues sending complete files. I can send an mp4 file where its source size is 3.85 MB but while being sent I get this error (image 4) and I can only stop the file transfer. When looking at the file that ended up after the transfer it is 3.949 Mb so it became slightly bigger somehow but it's unviewable.
2. When trying to interact with this mp4 file from #1, my Windows Explorer will stop responding and crash. this occurred on two different devices. I suspect maybe because the files are incomplete and that could be problematic or it could be in some other way related to problem #1

 

[joeschmuck]

I would have to say that your system appears to be stable based on those two tests.

Your NIC could be the issue. The NIC you have should be "Atheros AR8161L" and I personally have never heard of that before. That doesn't mean it's the issue but I suspect it is. If it were an Intel NIC, I wouldn't question it.

 

[Ibrahim05]

So I should purchase a better NIC (PCIE Ethernet card) which you presume could fix the issues rather than using the integrated motherboard ethernet port? If that's what you recommend I could try that. I'd assume this would also be a good way to potentially increase the speed of the server's data performance (in and out) by getting a card with more speed capabilities than the motherboard NIC, so I am not opposed to it.

Just want to make sure I understand what you're recommending, and curious if my assumption is correct or if this won't have an effect on performance.

 

[joeschmuck]

could fix the issues

Yes, it could. At some point we are making an educated guess.

getting a card with more speed capabilities than the motherboard NIC

Odds are you would get more speed with an Intel NIC over any other NIC because Intel processes the data in the NIC where other NIC cards often require the CPU to perform the data processing.

You do not need to spend a lot of money on a NIC. A very popular model (I have two myself) is EXPI9301CTBLK. I bought mine for $30 USD eleven years ago. You could purchase whichever model you desire however I would recommend you ask if it will work. I do not know of any Intel NIC that is not compatible with TrueNAS, but I do not know everything. The i210 chipset is found in many server motherboards these days, and of course intel keeps making newer models.

Again, this is just an educated guess because we have checked pretty much everything else.

Here is something else you can do to verify data transfers within your system... Make a duplicate copy of some of your data and see what happens. Hopefully all the data is transferred without issue.

WARNING !!! If you do not know what you are doing, then do not perform these steps if you do not have a copy of your data backed up. I have to assume you have some basic knowledge. While none of these steps will destroy your data, if you screw up then you could lose data.

ALSO, you need to have enough storage space to save this data.

0) Identify a dataset and directory that has a lot of small files (photos, music are common). And identify a directory which has several large files such as movies/videos. For the example here we will say dataset 'photos' is one, and dataset 'movies' is the other. Also for this example, your pool name is 'tank'.
1) Create a new dataset that you will be copying data to. I say to use a new dataset because cleanup will be very easy by just deleting the dataset. Let's call this dataset 'testarea'.
2) Now open up an SSH window or you could use the GUI Shell. I prefer the SSH window as you cannot accidentally close it as you could when using the GUI Shell.
3) Change directory to the testarea cd /mnt/tank/testarea and this should place you in the testarea location. Your cursor should clearly state this.
I am placing you in the destination directory because if you screwed up the command while being in the 'photos' directory, you could be making duplicates in your data area, that makes for a painful cleanup. Yes, I have done that before a long time ago, I will never forget it. I'm certain most of us have made that mistake, possibly several times.
4) Here is the first copy operation cp -rv /mnt/tank/photos . and you should be able to watch the files being copied.
5) If you still have the space you can copy the movies as follows...
7) Here is the copy operation, just like the last one cp -rv /mnt/tank/movies . and it should be copying files.
8) At any point during the copy operation you desire to stop, press CTRL+C several times, it will abort.

If all the files had no errors then transferring files internally is not the issue. This pretty much means it looks like the NIC is the issue. That is assuming your network is good and the computer you are transferring files with is not the problem.

Cleanup...
9) Let's cleanup our mess. Go back to the GUI and delete the dataset 'testarea'. That is it!

The 'cp' command, what are we doing? The 'cp' stands of course for 'copy'. The '-r' stands for recursive so all subdirectories are also copied. The 'v' portion is for Verbose so you can see something happening, otherwise the screen stays blank during the entire process and the small files can take a long time. We all like to see things working. I guess you could listen to the hard drives making noise but most of us like a visual feedback.

Alos, if you do not have a photos or movie directory, but you have some others that have lots of files, you can use those as well. If in doubt, ASK.

Okay, my fingers are tired.
Best of luck!

 

[Ibrahim05]

Sorry for the long gap in replies, I haven't forgotten lol

I don't have any photos or movies to transfer (because of the issue sending large files) so I just tried moving .txt files. I tried testing but it didn't seem to do anything. It looks like it says the usage of the commands is wrong. Changing directories was fine, but the verbose copying didn't work. 1. I tried as you said, 2. I tried without the verbose option, 3. I tried only a specific file without verbose, 4. I tried with their suggested syntax on a specific file. I'm not sure what could be the issue if I even use the suggested syntax, maybe you have some clues?

Regarding the NIC situation, I checked my NIC vs the one you recommended and they both seem to be gigabit, so if we aren't convinced beyond just a guess it could solve the issue I'll leave it as a backup solution if nothing else works. Since they are functionally identical I don't see how it would make a difference personally. The less CPU usage isn't important because the server doesn't seem to exceed 15% usage

I ran a little personal testing with theories I had and it seems that sending varied sized kb .txt files to the server and timing the transfer speed. Idk if that's helpful or not (if it is I can send some more info and observations) but basically it seems inconclusive.

1. 8kb can take ~2min + disconnection error to upload while a 9kb file can be done in ~15 sec. This makes me think along with the fact I keep getting disonnect errors while transferring that there might be some timing-related issues with these file transfers (maybe timing in relation to a DMA write error?)
2. So far, a file 24kb big will not transfer even if I wait ~6min and after 2 disconnects and reconnects.
3. File Explorer transfers are always stuck on the "calculating" phase and then when it's ready to send, the green progress chart zooms past in 2 seconds.
4. cpu usage is very low all the time. Even while transferring files it typically never exceeds 5% usage. It will always be at 0% and occasionally randomly increase to 2-3% for no observable reason.

Essentially the problems to fix are currently:
* the strange connection to the server that limits file transfers (so far we only know it occurs over the internet) with the error code "0x8007003B"
* Maybe related to the first problem, but there's a .png file in the server that I can't delete via File Explorer, every time I select or right-click it, File Explorer crashes. I know I could delete it via Shell, but maybe it gives a hint at the problem?

 

[joeschmuck]

maybe you have some clues?

The only thing I can think of is you are not a privileged users such as 'root'.

Regarding the NIC situation, I checked my NIC vs the one you recommended and they both seem to be gigabit, so if we aren't convinced beyond just a guess it could solve the issue I'll leave it as a backup solution if nothing else works. Since they are functionally identical I don't see how it would make a difference personally. The less CPU usage isn't important because the server doesn't seem to exceed 15% usage

RealTek has terrible support with TrueNAS, so it does make a huge difference, but your specific NIC I have never heard of so I have no idea if it's well supported or not. You cannot think that every piece of hardware has the same support, it does not work that way unfortunately. This is not an Apple world where hardware has to meet very strict requirements.

But if you are unable to move files around form the command line interface, you have larger problems. As I said, maybe it's a privilege issue but for now the NIC is on the back burner until you can figure out how to move files around.

Maybe TrueNAS just does not like your hardware at all.

Have you tried TrueNAS SCALE? It is build on Debian vice FreeBSD. It has different drivers, maybe this will work for you.