(HELP) HDD's making strange noise every now and then

[Johev]

Dear Community,

As always I come here for help whenever I had some issues with my FreeNAS box, which thus far has been minimal. I am still running the old 9.3 version, as I got errors updating every-time I've tried. The uptime has been more than 1 year, until I shut it down for maintenance.

My box which is giving my some HDD issues is the following: Dell T20 E3-1225v3, 12GB of ECC RAM, 1TB HDD (that came with the box) - it's used on it's own for TV recordings, 3*4TB Seagate surveillance drives in RAIDZ (store important files). I have scrubs set up every month and e-mail notifications if there are any issues. I've followed one of the Freenas #gurus tutorials for this (I think it's on the 28th day of each month).

Now recently my box started from time to time making noises (it sounds like if someone took a blade to the HDD while it's spinning and scratched it). The sound only lasts for a few seconds (max 2) and is a bit annoying , specially because I have it in my living room. It's very high pinched, thus impossible to ignore. I do not hear it during scrubs, just under normal idle operation. I tried to copy or read files from it and it seems to be working OK. I also tried to see the smart output and it's not throwing any errors. I don't know if it's the 1TB drive that came with the server, if it's any of the 4TB Seagate surveillance drives (which should still be under warranty) and therefore I am asking the community for assistance.

As mentioned I'm still running FreeNAS 9.3 (before the 9.1 came out), as I was never able to update it through the internal updater and therefore I've reverted to it.

Any ideas on what I should do?

 

[joeschmuck]

Are you running SMART Long Tests on your drives? If not then run it on all your drives. If you do then post the output for each drive of smartctl -a /dev/ada? and lets see if you have errors.

The most likely issue could be a failing fan in the unit creating this noise. This could be any fan, including inside the power supply.

 

[tvsjr]

Some drives also make odd noises when spinning up. Smartctl will tell the story if the drives are being spun down and spun back up.

 

[Johev]

Some drives also make odd noises when spinning up. Smartctl will tell the story if the drives are being spun down and spun back up.

I hear a constant noise of humming, thus I suppose that the drives are not being spun down too much. As for the command, could you please explain a little bit how to use it, or make it output to a file which I would then upload here? Many thanks.

 

[joeschmuck]

I would do the folowing to diagnose a failing fan...

1) open the case.
2) check to see if all fans are spinning, if not then you found the failure.
3) if all fans are spinning, shut down the box and unplug the power connectors to the hard drives.
4) power on the box and see if you can locate the failing fan.
5) repeat the power off/on sequence a few times (the reason for disconnecting the hard drives).

Typically it's obvious to locate a failing fan but ensure you look at all fans.

As for the command, could you please explain a little bit how to use it, or make it output to a file which I would then upload here?

In my signature is a link to the Hard Drive Troubleshooting Guide. But I'd open the case up and see if you can isolate where the noise is coming from.

 

[Johev]

@joeschmuck Many thanks for your quick reply. However I don't think it's a fan. I have some IT experience and it really sounds too high pinched to be a fan (I will nevertheless have a look at it). As for the Hard Drive troubleshooting guide, I've given it a quick read and I will be trying it out later today. Is there a way to output SMART data to a file which I could then copy to there? If yes, how could I do that? many thanks.

 

[joeschmuck]

If you use SSH program like Putty and then you can output all the data to a text file, this being the best way in my opinion. If you just want to type a command line thing then just pipe it to a file using the greater than symbol ">" like smartctl -a /dev/ada0 > /tmp/ada0.txt and then you would have that file. Like I said, I like using Putty but either method works fine.

EDIT: If the drives are making that noise, I'd be suprised if a smart long test would pass.

 

[Johev]

@joeschmuck Thank you so much for all the advice. I did some smart long tests, before I went to bed, however when I came back my laptop was suspended and therefore I could not find the result of the test. Is there somewhere I can retrieve the last long SMART test for the disks?

I have also tried the smartctl -a /dev/ada0 > /tmp/ada0.txt , however I don't know how to retrieve that file. I can't navigate with Windows Explorer to the USB drive where my Freenas is installed.

 

[danb35]

I have also tried the smartctl -a /dev/ada0 > /tmp/ada0.txt , however I don't know how to retrieve that file.

Rather than putting ada0.txt in /tmp/, put it in a shared directory on your pool. For example, if your pool is called pool, and you have a shared directory called share, run smartctl -a /dev/ada0 > /mnt/pool/share/ada0.txt. You should then be able to see ada0.txt in that shared directory on your client computers.

 

[joeschmuck]

when I came back my laptop was suspended

I thought a Dell T20 was a Mini Tower, not a laptop. We do not recommend using a laptop for FreeNAS.

Rather than putting ada0.txt in /tmp/, put it in a shared directory on your pool. For example, if your pool is called pool, and you have a shared directory called share, run smartctl -a /dev/ada0 > /mnt/pool/share/ada0.txt. You should then be able to see ada0.txt in that shared directory on your client computers.

This is true however it doesn't help the issue of if the laptop goes into sleep mode. Honestly I didn't think FreeNAS supported sleeping the entire system. I'm more conserned on if the hard drive will remain powered up long enough to complete a long test.

I have also tried the smartctl -a /dev/ada0 > /tmp/ada0.txt , however I don't know how to retrieve that file. I can't navigate with Windows Explorer to the USB drive where my Freenas is installed.

I have some IT experience

I don't think you have as much as you think. I don't mean to be a jerk here and I know I'm coming off like one but the stuff described above is very basic for any OS, unless you are an operator only. We really need to understand your knowledge level in order to provide the proper level of detail in the help information.

With all of that said, you can test a hard drive on a Windows system, get a program called smartmontools for windows. It opens a DOS command prompt and you type in a command like: smartctl -a c: > c:/drive_c.txt and the file will be in the directory specified. Change the drive letter c: after the -a to the drive you desire to look at or test. To run the long test replace the "-a" with "-t long". The long test will likely take several hours to complete, ensure the drive has power for the entire test. Once the test is complete you can command the drive results using the "-a" value. The drive retains the results of all the current tests so if you turn the unit off after the test is complete, you can pull the report anytime after.

Hope this points you in the right direction.

 

[danb35]

I thought a Dell T20 was a Mini Tower, not a laptop.

It is; I assumed that he was using the laptop to connect to the FreeNAS box--and in that case, the fact that it had gone to sleep is meaningless.

 

[Johev]

@joeschmuck I think that there must a be a misunderstanding. A Dell T20 is a tower server, I've also never seen a laptop where I could fit 4 x 3,5 inch drives in a RAIDZ :). My command line FreeBSD or Linux is not that great, however I have been building PC's with Windows systems for the past 20 years.

@danb35 Fully understood the issue, I was using putty on a laptop and it did not stay on long enough for the output of the smart long test.

I will try to run it again with saving the file on the shared directory of the pool. Thank you both for your assistance.

 

[Johev]

@joeschmuck @danb35 I have completed all the short SMART tests. Please find attached the txt files with the output. As for the long tests, I have tried the following command

Code:

smartctl -t long /dev/ada0 > /mnt/pool/share/ada0.txt

(with the adaptation of my volume and share name). However I was not able to get the output recorded, as the output was basically the following for all of them.

smartctl 6.3 2014-07-26 r3976 [FreeBSD 9.3-RELEASE-p28 amd64] (local build)
Copyright (C) 2002-14, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF OFFLINE IMMEDIATE AND SELF-TEST SECTION ===
Sending command: "Execute SMART Extended self-test routine immediately in off-line mode".
Drive command "Execute SMART Extended self-test routine immediately in off-line mode" successful.
Testing has begun.
Please wait 498 minutes for test to complete.
Test will complete after Sun Jul 23 08:58:46 2017

Use smartctl -X to abort test.

Is there a way to read the last SMART long test output, or get the output of the long test recorded? This way I would not have to run it all again :).

 

[danb35]

Is there a way to read the last SMART long test output, or get the output of the long test recorded?

When you start a SMART test with smartctl -t (type) /dev/whatever, you're starting a test that's internal to the drive. The smartctl program should return you almost immediately to the shell prompt, with a message like what you posted above. After the amount of time it specifies, you can see the results, along with the rest of the SMART data, by running smartctl -a /dev/whatever. Ideally, you'd then copy and paste that output into a post here, using code tags (not command tags). Failing that, you'd redirect it to a file ( smartctl -a /dev/whatever > /mnt/pool/share/smart.txt), and then copy and paste it into a post here, using code tags (not command tags).