CRITICAL: The volume Volume2 (ZFS) status is DEGRADED help

[AirNewZealand15]

Hello

I have been using FreeNas for about 6 months now and have just been alerted to: CRITICAL: The volume Volume2 (ZFS) status is DEGRADED.

What is the best way to rebuild my raid array? I don't believe the drive(s) are faulty.

Any instructions would be much appreciated. I have googled it and also had a poke around the freenas GUI but just can't seem to see where to rebuild the array.

I am running FreeNas 8.0.4-Release-x64.

Thanks in advance.

 

[survive]

Hi AirNewZealand15,

Take a look at this:

http://www.freenas.org/images/resou...8.0.3_guide.html#__RefHeading__9510_322745209

There isn't really a way to just rebuild the pool.....something has convinced zfs that the drive is dodgy and you are going to have to pull the drive (good idea to test it with something to check the smart report) and either replace it or wipe it to get the pool healthy again.

-Will

 

[AirNewZealand15]

thanks for the prompt reply!



I would assume from the image above that the second drive is the 'faulty' one as it has no name etc

Also, under my SMART tests I have 'no SMART tests defined' so I assume that I have to add them obviously. What is the best test(s) to add here?

THANKS!

 

[paleoN]

From a SSH session as root paste the output of:

Code:

zpool status -v

camcontrol devlist

gpart show

To view the SMART info:

Code:

smartctl -q noserial -a /dev/adaX

 

[Stephens]

I don't believe the drive(s) are faulty.

Absent any evidence about why you believe this (what tests you've run to verify it), a more accurate word than "believe" might be "hope". Definitely run the smartctl test paleon referenced and report back with the output.

 

[AirNewZealand15]

Absent any evidence about why you believe this (what tests you've run to verify it), a more accurate word than "believe" might be "hope". Definitely run the smartctl test paleon referenced and report back with the output.

Good call. I have NOTHING to backup my statement about the drives not being faulty except the fact that they are quite new, however as you know even new drives can fail.

PaleoN - I don't connect to the freenas via SSH, I'm a bit of a n00b to it and all the setup I have done is via the browser gui - can I run your suggested test via the gui? If not, I will just remove both the drives and pop them into my PC to run the SMART tests.

Thanks.

 

[paleoN]

PaleoN - I don't connect to the freenas via SSH, I'm a bit of a n00b to it and all the setup I have done is via the browser gui - can I run your suggested test via the gui?

Not on 8.0.4. Take the whole one and a half minutes and setup SSH. Personally, I use PUTTY as a client. Technically you can use the shell directly on the FreeNAS box. It's just hard to copy & paste.

If not, I will just remove both the drives and pop them into my PC to run the SMART tests.

The smartctl command I gave views the SMART info and doesn't test the drive. If one or more of the drives are failing, running any type of tests on them can cause them to fail sooner.

I wouldn't remove the drives just yet unless you don't mind losing the zpool. It can also make troubleshooting this harder.

 

[AirNewZealand15]

Ok I removed both the drives and tested in my PC (PaleoN this was before your reply) and one drive does in mac have issues with the SMART:



So I guess I will just get that replaced under warranty.

Stangely though, when I put the drives back into the FreeNas and booted it up, it now reports my RAID array as healthy!

 

[joshg678]

Scrub it.
Get that drive replaced afterwords.

 

[Stephens]

Ok I removed both the drives and tested in my PC (PaleoN this was before your reply) and one drive does in mac have issues with the SMART:

I'm glad you took my reply about in the way it was meant. I know where you're coming from because why in the world should a new drive be bad after we've just purchased it, waited for it, installed it, etc. But I've been doing SMART long tests on my drives as soon as I get them because of how generally unreliable (anecdotally) larger consumer drives are becoming and I'm seeing what I'd call DOA's -- drives failing SMART tests right from the initial bootup. I personally recently had to RMA a brand new drive on my current NAS build. Brand new, as in I plugged it in, then ran SeaTools. And it failed.

FYI, I didn't do my testing in FreeNAS at all. I made a usb flash of the Ultimate Boot CD, then booted that, went into HDD tools, Diagnosis, then "SeaTools" (the newer version). That not only has a graphical interface, but it creates a log you can copy back to your usb flash drive. I included this for Seagate along with the RMA'd drive.

 

[AirNewZealand15]

Thanks Stephens, that's good to know.

My FreeNas has now corrupted itself (it appears). I logged in to check the raid array and clicked on volumes but it just wouldn't load up volumes at all. So I restarted the FN through the GUI and after that is now says geometry mismatch on my 2gb flash drive that I had freenas loaded on and booting from. So tomorrow morning I am going to reinstall freenas and then restore my settings from the backup I made and I'm assuming (hopefully) all my volumes will be there and I won't lose and data... Right??

My faulty 2TB drive has been sent back to the supplier today so my raid array was only running with one drive - but I thought this would be fine - it wouldn't have caused the corruption right?

 

[Stephens]

I believe geometry errors on the boot flash drive are "normal". It's possible you always got that message but never noticed it before. I get it.

I don't know if you followed the documentation procedure to remove the bad 2TB drive or if you just yanked it out of your system. But yes, you should be OK. Drop to a command shell and do a "zpool status". You should see something like "DEGRADED" but OK.

I can't think of any reason your data should be at risk other than if the remaining drive goes bad. Just make sure you read the documentation for anything you plan to do. It's not a good environment for just trying things to see how they work out.

 

[survive]

Hi AirNewZealand15,

Go out & get yourself a bigger flash drive....most 2GB keys are just a hair too small for the FreeNAS image.

-Will

 

[joshg678]

+1 on the USB being too small.
I plan to purchase a 4GB for mine.

I Believe FreeNAS partitions it to 50%/50% regardless of the size, this way you have the ability to upgrade with larger images in the future if need be.

 

[paleoN]

I Believe FreeNAS partitions it to 50%/50% regardless of the size, this way you have the ability to upgrade with larger images in the future if need be.

This is incorrect. FreeNAS creates 4 partitions, the first about two 940M & the last two are fairly small. The rest of the drive is unused.

 

[AirNewZealand15]

I believe geometry errors on the boot flash drive are "normal". It's possible you always got that message but never noticed it before. I get it.

I don't know if you followed the documentation procedure to remove the bad 2TB drive or if you just yanked it out of your system. But yes, you should be OK. Drop to a command shell and do a "zpool status". You should see something like "DEGRADED" but OK.

I can't think of any reason your data should be at risk other than if the remaining drive goes bad. Just make sure you read the documentation for anything you plan to do. It's not a good environment for just trying things to see how they work out.

Ok . I will get a new flash drive today. I should have mentioned that my freenas now DOES NOT BOOT and that is why I was mentioning the geometry message because that is where it stops. Check out my photos:



Main messages are: root mount failed, startup aborted. But please have a look at those photos

Sadly I did not read any documentation about removing drives... I just rushed into it. My Freenas has been running perfectly up until I removed that drive and now it won't boot... seems a bit more than a coincidence right?

 

[joshg678]

This is incorrect. FreeNAS creates 4 partitions, the first about two 940M & the last two are fairly small. The rest of the drive is unused.

Thanks for clarifying it. I thought that seemed different from what i had remembered when i read it somewhere.

 

[AirNewZealand15]

seems like I should have followed this and marked the faulty disk as offline. doh.

 

[AirNewZealand15]

Ok I am all back up and running now. I reintalled FN on an 8GB flash drive and then restore my settings from my backup I made a while ago. all my shared work fine WOOHOO

BUT - I now want to mark the removed drive as offline... and the documentation says "go to volume manager" - sounds like a newer version of FN? What is the process for 8.0.4? Or should I just leave it and when my replacement drive arrived I'll just connect that in and off it goes?

 

[Stephens]

If you go to the documentation index page, it tells you which version the documentation it is for (if there are any further releases, changes are noted in the release notes). Also, older versions of the documentation are still out there. I think the equivalent instructions for 8.0.4 (8.0.3 really) are here.