Advice on 'Current Pending Sector'

[Jacko_]

Would someone be able to tell me if my HDD is in need of being replaced or if this is something that can be fixed?  I think i caused this issue when i pulled a drive by mistake.  I've looked through the SMART test but i don't really know what i'm looking for - would be good to understand the issue and resolution.

 

I'm wondering if it's worth replacing the drive and then zero fill formatting this one.  I don't think it's a HW issue with the drive - i think it's my fault by removing the drive.  Doh.

 

Many thanks

 

tower-smart-20220825-0939.zip tower-diagnostics-20220825-0942.zip

Edited August 25, 2022 by Jacko_

[JorgeB]

You can run a SMART test to confirm if they are a problem of not, if they are filling the drive with zeroes might remap them.

[Jacko_]

Hi JorgeB

 

I attached the SMART results in the previous message.  It looks to be passing the test i don't know if i'm missing something i should be reviewing and taking note of.

[JorgeB]

Forgot to specify, run the extended test, short one doesn't do a surface scan.

[JorgeB]

Make sure disk spin down is disabled before starting.

[Jacko_]

I thought i did an extended test but i must have hit the wrong button.  I've attached the extended test results.

tower-smart-20220825-1309.zip

[JorgeB]

Test failed, either replace the disk or try a full disk write to see it the sectors are remapped and don't come back again.

[Frank1940]

@Jacko_, what I have done in this situation to replace disks like this with a spare--- using Unraid's ability to rebuild the data on the replacement.  Then, I will run (at least) two preclear cycles (Using either the plugin or the Docker) on the disk to see what the SMART results are on the disk after each cycle.  (You could also use the disk manufacturer's test suite to do the same thing.)

 

PS---  I also wait until the rebuild is complete before I start any testing.  (Most of the time, only one or two files will have problems so that 'damaged' disk is a backup of most of the data...)  I also have cold spares (the same capacity as the parity drive) for quick replacement of any drive with any issue.  When I first purchase a cold spare, I immediately run enough preclear cycles on it to put 70-100 hours of intensive disk activity on it to make sure that it will not be a victim of 'infant mortality'.   (Again, you could use the manufacturer's software to do the same thing.)  You could lose a good portion of the warranty this way but I feel that if the disk survived those first hours of being hammered by the testing, it will more than likely last through the warranty period. 

 

If the disk's testing shows that the disk is not a potential failure waiting to happen, I will make it the new cold spare.  (Often, the 'disabled disk' --- ones with the red "X" --- will test as good indicating that something else caused the write operation to fail...)

[Jacko_]

Hi both,

 

Thanks for input on this. I'll need to figure out how to remove copy the data off the drive so i'l do some research on that.

[itimpi]

2 minutes ago, Jacko_ said:

I'll need to figure out how to remove copy the data off the drive so i'l do some research on that.

Since you are on Unraid 6.1.3 then installing the Dynamix File Manager plugin and using that is probably the easiest and least error prone way to do this.

[bonienl]

The Dynamix file manager has an easy way built-in to move content from one disk to another disk (provided there is enough space on the destination disk)

 

1. Open the file manager from the top-right icon

 

2. Select the Move operation from the source disk (e.g. disk1)

 

3. Select the target disk to move to (e.g. disk4)

 

4. Select PROCEED and all content of disk1 will be moved to disk4 (existing files will not be overwritten unless explicitly selected)

 

5. You can follow the progress in the opened window or close it (X in top-right) and let it run in the background.

 

[Jacko_]

Thanks both, that certainly looks easiest route.  New disk is installed and its currently clearing so hopefully i'll be able to get the 'failing' drive out of the pool and cleared successfully.   What's a bit strange is i couldn't use the disk preclear to do a first pass on the disk before adding it to the pool - it didn't like the MBR sig for whatever reason.

 

So rather mess about i just added it to the pool and unraid is clearing the disk - kind of like the belt and braces approach but hopefully this is enough.  I just hope i can pre-clear the old drive using the plugin before re-adding it to the pool if successfully passes.

 

Wish me luck.  

 

 

 

 

 

 

####################################################################################################

#                    Unraid Server: verifying Preclear State of disk  5QJLTU0B.                    #

#                      Verifying disk '5QJLTU0B' for Unraid's Preclear State.                      #

#                                                                                                  #

#   Step 1 of 2 - Verifying Unraid's signature:                                    FAIL            #

#                                                                                                  #

#                                                                                                  #

#                                                                                                  #

#                                                                                                  #

#                                                                                                  #

#                                                                                                  #

#                                                                                                  #

####################################################################################################

#       Cycle elapsed time: 0:00:01 | Total elapsed time: 0:00:01                                  #

####################################################################################################

 

####################################################################################################

#   S.M.A.R.T. Status (device type: default)                                                       #

#                                                                                                  #

#   ATTRIBUTE                   INITIAL STATUS                                                     #

#   Reallocated_Sector_Ct       0       -                                                          #

#   Power_On_Hours              0       -                                                          #

#   Temperature_Celsius         32      -                                                          #

#   Reallocated_Event_Count     0       -                                                          #

#   Current_Pending_Sector      0       -                                                          #

#   Offline_Uncorrectable       0       -                                                          #

#   UDMA_CRC_Error_Count        0       -                                                          #

#                                                                                                  #

#                                                                                                  #

####################################################################################################

#                                                                                                  #

####################################################################################################

--> RESULT: FAIL! /dev/sdl DOESN'T have a valid Unraid MBR signature!!!

 

[Jacko_]

Well this has been a seamless process so far.

 

Interestingly, when i stopped the array to remove the offending disk i removed the partition and could pre-clear (or at least kick it off) without issue.  Something strange happening with MBR on the new drive that meant it couldn't be pre-cleared for whatever reason.

 

So now the long wait and hopefully no failures along the way with this older drive and it passes the test and i can re-introduce it to the array.  Failing that, maybe i need to see if i can RMA it but it's a shucked drive so no idea which serial number caddy it came out of.

 

I have this little green tick where Disk 3 used to be.  In the file manager i can delete the drive but will it re-order the remaining drives keeping them sequential?  It will annoy me if it doesn't.  I could wait to see if the drive passes pre-clear and reintroduce at Disk 3 i guess...

 

[Frank1940]

5 minutes ago, Jacko_ said:

i need to see if i can RMA it but it's a shucked drive so no idea which serial number caddy it came out of

 

I seem to recall that some folks have used the serial number on the drive for the RMA process.   This policy may vary from manufacturer to manufacturer.  The second problem is that the 'date of purchase' becomes what the manufacturer says it is as that date will precede the date you purchased the drive. 

[itimpi]

13 minutes ago, Jacko_ said:

I have this little green tick where Disk 3 used to be.  In the file manager i can delete the drive but will it re-order the remaining drives keeping them sequential?  It will annoy me if it doesn't.  I could wait to see if the drive passes pre-clear and reintroduce at Disk 3 i guess

At the moment Unraid should be emulating disk3 using the combination of the other disks plus parity to act as if it were still present.  If you replace the missing disk3 then Unraid will rebuild the contents of the 'emulated' disk3 on to this replacement.  If you try to do anything else this will invalidate parity (and parity will need to be rebuilt) and thus the contents of disk3 become irretrievably lost so be aware of this.

[Jacko_]

15 minutes ago, Frank1940 said:

 

I seem to recall that some folks have used the serial number on the drive for the RMA process.   This policy may vary from manufacturer to manufacturer.  The second problem is that the 'date of purchase' becomes what the manufacturer says it is as that date will precede the date you purchased the drive. 

 

Interesting - i'll do some digging on the WD site to see what it says.

 

8 minutes ago, itimpi said:

At the moment Unraid should be emulating disk3 using the combination of the other disks plus parity to act as if it were still present.  If you replace the missing disk3 then Unraid will rebuild the contents of the 'emulated' disk3 on to this replacement.  If you try to do anything else this will invalidate parity (and parity will need to be rebuilt) and thus the contents of disk3 become irretrievably lost so be aware of this.

 

Ah ok.  So currently i have used file manager to move data to another drive.  I then stopped the array, removed the offending drive (disk 3) and restarted the array, so the data now exists on a new disk (disk 6).  In this scenario, what are my options without the array trying to rebuild from parity?

[Frank1940]

6 minutes ago, Jacko_ said:

Interesting - i'll do some digging on the WD site to see what it says.

 

https://support-en.wd.com/app/warrantystatusweb

[Jacko_]

 

Just my luck. 😆

 

I guess i have to keep those fingers crossed it passes the pre-clear.  I guess this is where it pays you to buy proper drives in the first place - longer warranty and better endurance.

[Jacko_]

Looking promising, but only 7% through pre-clear.

 

[itimpi]

21 minutes ago, Jacko_ said:

In this scenario, what are my options without the array trying to rebuild from parity?

You can use the New Config tool to change the assignments to what you want  and then rebuild parity based on those assignments.

[Frank1940]

1 minute ago, Jacko_ said:

 

Just my luck. 😆

 

I guess i have to keep those fingers crossed it passes the pre-clear.  I guess this is where it pays you to buy proper drives in the first place - longer warranty and better endurance.

This is usually the case of the drive being out of warranty.   The manufacturers have a tendency to use the USB house drives as a means to move old inventory.  They can put any model drive into them!  The consumer is never given the information.

 

But think about it a bit.  How many drives have you had fail while they are under warranty?  Now compare this number to the number of drives you have purchased over the years.   How figure how much you saved by buying that schucked drive vs the fully-warranted drive.  Divide cost of fully-warranted by savings.  Then divide the number-of-drives-purchased by the number-of-RMA-drives.   Would you be ahead of the game?    (As far as I am concerned, the warranty on most things is a nice bit of protection if the thing is basically broke-out-of-the-box.  Never purchase an extended unless you know it will absolutely fail before the warranty is up!)