How To Determine Which Hard Drive Failed?

[Clinderw]

Hello,

I have 10 hard drives, most of which have similar model numbers or makes. In preparation for one dies, how will I know which one is failing (or failed)? What I need to avoid is pulling a working hard drive from the ZFS1 set up and have the RAID be unrescueable.

I have SMART alerts set up with daily alerts letting me know the health of the drives/system but if one of them did go down I wouldn't be able to determine which one did based on that information.

Thank you in advance for the assistance.

Chris

 

[Z300M]

Hello,

I have 10 hard drives, most of which have similar model numbers or makes. In preparation for one dies, how will I know which one is failing (or failed)? What I need to avoid is pulling a working hard drive from the ZFS1 set up and have the RAID be unrescueable.

I have SMART alerts set up with daily alerts letting me know the health of the drives/system but if one of them did go down I wouldn't be able to determine which one did based on that information

At Storage -> View Disks you will see a column headed "Serial" with the Serial No. of each disk. Depending on how your disks are mounted physically in the case, you can either mark the Serial No. on each drive so it's readable without removing the drive, or keep a list showing what drive is mounted where.

BTW, I would not set up an array of ten drives in RaidZ1: RaidZ2 would give you a much better safety margin. I am currently running just four disks in RaidZ1 but am planning to destroy that pool and replace it by one using RaidZ2.

 

[gpsguy]

Here's a second vote for putting labels with the serial numbers on the end of the drive.

Sometimes, while troubleshooting, users end up moving drives around and then end up confused and pull the wrong drive.

ADepending on how your disks are mounted physically in the case, you can either mark the Serial No. on each drive so it's readable without removing the drive....

I was waiting on our friend cyberjock to make a comment about RAIDz1. With that many drives, if plan to rebuild the array, consider going to RAIDz3.

 

[cyberjock]

Yeah..gpsguy knows my mantra too well. RAIDZ1 is a recipe for disaster.. go RAIDZ2 or RAIDZ3.

 

[Clinderw]

Thanks everyone for the responses - I think i'll take the advice of labeling the drives.

I did RAIDZ1 because i figured the odds of two failing are slim to none and i have a spare readily available in the box. My set up is 5X3tb (Vdev1) + 4X2tb (vdev2) + spare 3tb over the pool. Knowing this would you still recommend RAIDZ2? I am at the point now with my data consumption that i could easily make the switch but another month or two would make this a very unpleasant process. I did a decent amount of homework but no thread seemed to make a convincing enough story for me.

Thanks in advance

Chris

 

[cyberjock]

Heck yeah. Did you read the link in my sig(the one that RAID5/RAIDZ1 was declared unreliable... in 2009)?

 

[Clinderw]

Thanks - just saw that. I will begin the transition this evening. Thanks for the prompt responses!

Is there anything i need to do to put this thread in solved or finished status?

 

[KevinM]

Thanks - just saw that. I will begin the transition this evening. Thanks for the prompt responses!

Is there anything i need to do to put this thread in solved or finished status?

FWIW, what I did was write down the serial number of each disk as I added them to the enclosure, then in Storage > View Disks I edited each drive and added the chassis location. For example, disk da0 with serial Z2963S470000C3262PJ5 has the description 3T SAS: Back, Row 3, Col 1. This is for a Supermicro e1r36n which has disk enclosures on the front and back of the chassis.

I also second the recommendation for RAIDZ2. Generally the recommendation for sizing is 2n+2, so 8 data drives + 2 parity drives = 10 total would work well with your setup.

 

[Subwoofer Jesus]

I have taken all the advice here to heart THANK YOU I am building a FreeNAS and learning on the way,\.
I currently have 6 250 gigs as a test bed to learn.

I set it up with RAID Z2 after a lot of reading and discussion on this site when one drive kept getting unrecoverable sector errors
and guess what, I did not label them so I did it the hard way. disconnecting each drive until I found the one. LOL
now I label and found I enjoy Labeling :)

now trying to figure out if is one of the drives or a port on the card :)
More reading and testing Thanks once again to the Forum here

 

[Clinderw]

two of my drives come back with no serial number - is there a script that i can run to get to this information?

Any idea why that would be?

thanks,
Chris

 

[Yatti420]

two of my drives come back with no serial number - is there a script that i can run to get to this information?

Any idea why that would be?

View attachment 4782

thanks,
Chris

How are they connected? If you run smartctl -a /dev/da# do you get the info? If you have a raid card or are using a raid controller that's probably why your not getting that info..

 

[Clinderw]

Yep when i run the command it comes back with info - one is a seagate and one is a western digital. Both are their NAS drives.

I am using a IBM M1015 (IT) controller but i'm not sure if these two drives are hooked up through it. There are 6 drives hooked up through the controller and only 2 drives in my system are returning blanks.

They are either through the controller or the motherboard - i have more than 2 drives hooked up to each.

This got me what i needed in the meantime, seems like a bug though.

Thanks,
Chris

 

[Yatti420]

hmm odd.. Should be fine.. If you can see it via smart maybe just a GUI bug..

 

[William Bravin]

At Storage -> View Disks you will see a column headed "Serial" with the Serial No. of each disk. Depending on how your disks are mounted physically in the case, you can either mark the Serial No. on each drive so it's readable without removing the drive, or keep a list showing what drive is mounted where.

BTW, I would not set up an array of ten drives in RaidZ1: RaidZ2 would give you a much better safety margin. I am currently running just four disks in RaidZ1 but am planning to destroy that pool and replace it by one using RaidZ2.

Hi all I'm new to freenas and i just completed the installation. Freenas sees all my drive an it tells me in the alerts that one or more devices are degraded. under storage or in dashboard i cannot see which drive has the issue and all the lights of the drived on my NAS are green. I have a pool 4 X 2 tb and 4 by 1 tb drives. Yet i only have 7.53 available as disk space.

How do i identify the bad drive or cable?

thank you for your help

 

[gpsguy]

Run zpool status and note the bad gptid.

Using @Bidule0hm 's script to "Display drives identification infos (Device, GPTID, Serial)" - https://forums.freenas.org/index.ph...tion-and-backup-the-config.27365/#post-175548 you can map the gptid to the serial number.

How do i identify the bad drive or cable?

How much disk space you have, is dependent on how you laid out your pool. If you post the output of zpool status in code tags, we can see how you did it.

We'd prefer that you had started a new thread, vs necroing a 4 year old one. But, we do thank you for trying to search for an answer.

 

[William Bravin]

Run zpool status and note the bad gptid.

Using @Bidule0hm 's script to "Display drives identification infos (Device, GPTID, Serial)" - https://forums.freenas.org/index.php?threads/scripts-to-report-smart-zpool-and-ups-status-hdd-cpu-t°-hdd-identification-and-backup-the-config.27365/#post-175548 you can map the gptid to the serial number.



How much disk space you have, is dependent on how you laid out your pool. If you post the output of zpool status in code tags, we can see how you did it.

We'd prefer that you had started a new thread, vs necroing a 4 year old one. But, we do thank you for trying to search for an answer.

Thank you for this
I looked at the status as you mentioned and I found a drive that was degraded. I removed it and will be purchasing a new one. All other are fine and now My available space has gone up to 10.75 TB. Once i have the drive i will inset it and this will tell me if it is the drive or a bad cable. or controller card

Thanks once again

Because of scree size the 3 dots were not visible may i suggest the the "pool status "option be set as a sub-menu as this would be useful for all pools,

 

[gpsguy]

Did you follow the instructions in the documentation - Replacing a failed drive - http://doc.freenas.org/11/storage.html#replacing-a-failed-drive before removing the drive?

Assuming you were using some form of redundancy, ie. RAIDz or mirroring, removing a drive shouldn't change the available space.

If you can't upload the results from zpool status in code tags, could you take a picture and post it here?

I removed it

 

[danb35]

the 3 dots were not visible

Three dots? So you're using the beta? That would have been a good thing to have mentioned in your post. And really, it's a good idea to mention your version in any post where you're asking for help.