Drive Failing/Failed - Options?

[Connor123]

Hey

 

Just logged onto the GUI to find a data drive is failing/failed due to Reallocated sector count (see image) and also attached SMART report.

Unraid GUI is flagging this drive as disabled and content emulated.

 

1st question is: could this be getting reported in error? 2 weeks ago I added a new drive and these errors have only just come up - could this be a sata cable issue from when I was adding the new dive?

 

Otherwise - assuming the drive is failed and it's reporting correctly:

 

From my quick look online I should take this as dead and get it RMA'd

 

 

I'm now trying to figure out best method to get this drive out of the array given the following information:

• My 6 SATA ports are all being used.
• I have 1 parity drive
• 2 weeks ago I pre-cleared and added 1 data drive to the array- it is still empty.

 

 

With regards to the last point - do I have an option to remove that empty drive from the array, keep parity and then follow the normal process to replace a failing drive? 

Thinking about it I'm not sure if that is any better than just using unbalance to transfer from the emulated drive to this newer drive & then removing the failed one from the array?

 

 

Any help appreciated!!

 

captain-smart-20200208-1754.zip

[itimpi]

The easiest thing would be to replace the failed drive with a new one and rebuild it.     However if you are not going to have a drive available to that then your best bet is as you say to transfer the contents of the emulated to the empty drive in the array.     When that completes you can use the Tools->New Config option to reset the array to not include the failed drive, and rebuild parity to match the new configuration.   Until that all completes you will be running without protection against another drive failing.

[Connor123]

Great - thanks itimpi - so steps are

1. Keep all drives in array as is.
2. Unbalance failed drive/emulated drive to other disk/s
3. Tools - new config & drive out of action for ~1 day whilst parity rebuilds.

Is there no option to remove the recently added drive and keep the array in parity and do something more akin to a quick swap? Sorry - first time doing these

 

If I were to get a replacement drive quickly would I follow the same above steps except after unbalance I would take the array offline and turn off system, replace the failed drive with the new one and then on start-up go with the new config option and rebuild parity from scratch?

 

 

Other question not quite Unraid related.

I can send this failed drive back to amazon to get refund and just purchase a new one or I can RMA it via WD.

Is either of these more preferable? I read that WD would only provide refurbished via RMA.

 

 

 

Edited February 8, 2020 by Connor123

[itimpi]

5 minutes ago, Connor123 said:

Great - thanks itimpi - so steps are

1. Keep all drives in array as is.
2. Unbalance failed drive/emulated drive to other disk/s
3. Tools - new config & drive out of action for ~1 day whilst parity rebuilds.

Is there no option to remove the recently added drive and keep the array in parity and do something more akin to a quick swap? Sorry - first time doing these

 

 

Other question not quite Unraid related.

I can send this failed drive back to amazon to get refund and just purchase a new one or I can RMA it via WD.

Is either of these more preferable? I read that WD would only provide refurbished via RMA.

 

it is worth pointing out that you can continue to use the array while parity is being rebuilt (although performance may well be degraded).
 

The problem with removing a drive is that you cannot remove a drive without affecting parity unless it is all zeroes.    The moment you format a drive to create an empty file system on it this is no longer the case,   What you want could theoretically be done using the Alternative method on This link.  however it does not seem to offer any advantage over the method previously discussed.

 

it is really up to you which route you follow for a replacement.   If Amazon will replace it then you will get a new drive.   However depending on how long you have had it they may well tell you to take the RMA route anyway.   If you get a replacement drive then you could simply rebuild onto the replacement without using Unbalance to clear the emulated drive as the rebuild process simply ends up putting whatever is showing on the emulated drive onto the replacement.

[Connor123]

Thanks @itimpi - massive help.

 

I will go ahead with:

1. Unbalance copy files from emulated to other disk.
2. New config and rebuild parity without failed disk included in array
3. Send failed disk back to Amazon (they have already agreed to process refund)
4. Purchase new disk and do the normal pre-clear and add to array.

 

Only problem I have with "If you get a replacement drive then you could simply rebuild onto the replacement" is my sata ports are full so I don't think that is an option?

[trurl]

12 minutes ago, Connor123 said:

Only problem I have with "If you get a replacement drive then you could simply rebuild onto the replacement" is my sata ports are full so I don't think that is an option?

You can remove the original drive and use its connection for the new drive.. That original drive is not being used now (disabled and emulated) and it is not used during rebuild.

[Connor123]

6 minutes ago, trurl said:

You can remove the original drive and use its connection for the new drive.. That original drive is not being used now (disabled and emulated) and it is not used during rebuild.

Understood - my bad! 

 

Thanks both for you help!!

[itimpi]

I would also suggest you do not add an additional drive to the array if you do not need it yet.   It just adds a potential additional point of failure.  You might want to just leave it on the shelf until needed.

[Connor123]

Fair point - in which case would pre-clearing it when it arrives to make sure it's OK and then letting it just sit there outside of the array un-used be a good enough option?

[itimpi]

1 minute ago, Connor123 said:

Fair point - in which case would pre-clearing it when it arrives to make sure it's OK and then letting it just sit there outside of the array un-used be a good enough option?

Pre-clearing it can be a good idea as that speeds up adding it to a new drive slot.   However I would avoid plugging it in at all if not needed unless there Is a reason such as not being able to easily access the server.