How To Fix: Degraded Pool

by ·

…it’s not the end of the world. Not yet.

It was a very nice day, when just out of blue, 2 of my 6 hard drives in RaidZ2 configured pool, start having hiccups (read: S.M.A.R.T errors). A SMART error is not the end of the world and it’s more like warning signs than an actual problem. More often, it will give you plenty of time to replace playing up disk(s).

So far, I’ve found that first of my Seagate 7200.11, 500GB – dev/ada4 with 13690 working hours has 3 different problems, each qualifying disk for replacement.

  1. Reallocated_Sector_Ct – 10
  2. Spin Retry – 2 (I could live with this one)
  3. Reported_Uncorrect – 530

Also, second Seagate 7200.11 with 8490 working hours has its own quirks:

  1. Spin Retry – 51
  2. Reported_Uncorrect – 33
  3. UDMA_CRC_Error_Count – 280 (It could be just faulty cable)

Again, not the end of this hard drive yet, but… you need to think about new one as well. Fairly quickly.

Today, arrived 1st replacement (I will never buy more than 1 disk from the same batch): Toshiba DT01ACA200 – 2TB Hitachi debranded hardware with 2 years warranty Return To Base. One thing to remember with Toshiba is that in case of damaged drive, you have to deal with seller and NOT Toshiba directly as they don’t have returns facility (at least in the UK)

Drive arrived from DABS.com in nice small parcel via Signed Delivery. Disk well packed, with nice, big FRAGILE  letters (shame, that it was inside the grey plastic bag, but… never mind). Inside, my Toshiba in the antistatic sealed plastic bag.

To check 2TB, I had to run the EXTENDED SMART TEST, via any S.M.A.R.T. monitor tool (I used well tested Hirens boot 15.2 from here) and it took over 4 hours to complete. Once completed, it was ready to go in my NAS4Free box. It was way easier for me to use my other media box to connect this hard drive and run test, than actually put in my NAS, as I have too many securing screws 😉 but you can feel free to just replace old HDD with the new one in your NAS4Free box.

How to replace damaged drive in ZFS? Actually, it’s very easy:

  1. Once you have physically replaced your old drive with new one.
  2. Start N4F and wait until it is up and running fully.
  3. Login into web and go to DISKS->Management
  4. Press Clear Config and Import Disks
  5. You should see new hard drive – in my case it was /dev/ada4 – TOSHIBA DT01ACA200, still not part of ZFS my pool
  6. You can use SSH, or pick menu ADVANCE->Command or like me, just switch monitor from DVI to VGA to see output from N4F box and choose menu 6 – Shell
  7. Type (in my case) zpool replace RaidZ2 /dev/ada4, where RaidZ2 is your pool name, /dev/ada4 is the number of your new drive, in my case 4, see point no. 5)
  8. If you type zpool status RaidZ2, you should see that your zpool is re-silvering, which mean replacing missing data.
  9. When it’s finished (timewise will depend on how much data you have and speed of your hard drive, for me 300GB was about 15 minutes), you should see your zpool status as ONLINE. No more degraded.

As you can see, I’ve replaced 500GB with 2TB, but my zpool is still worth of 4x 500GB. Until all disks are not 2TB, my zpool will not grow. Once all are 2TB, I will have more or less 4 x 2TB, usable around 7.2TB. Enough for a long time… At least for me 😉

Tags:

You may also like...