Too many wrong and/or missing disks! v5.0.5

[AdamEdmonds]

Hello all.  I'm encountering the error message Too many wrong and/or missing disks!.  My current situation is a I have 1 known bad drive disk2 (was a 1.5Tb clicking and other mechanical noises).  I was attempting to replace that drive when I encounter the error message mentioned above.  Disk3 is reporting Smart errors and I fear it may be dead.  

I have check, re-check, and tripple checked power and data (at drive and at Motherboard) for all drives.  Most of the posts I've read about this ask for diagnostics, but I'm unsure how to get  this information on my version. 

The attached image is what I'm seeing on the main page, but I'm unable to start the array under either of these scenarios:

1) Without disk2

2) Attempting to replace disk2 with new drive.

3) Without disk2 and disk3

 

Any help or guidance would be appreciated.

[trurl]

17 minutes ago, AdamEdmonds said:

Most of the posts I've read about this ask for diagnostics, but I'm unsure how to get  this information on my version. 

There is no way to obtain the diagnostics mentioned in other threads for this very old version of Unraid.

 

I have moved your post to the Legacy Support subforum. See this post pinned near the top of this same subforum for instructions on how to get us the syslog and SMART for all disks:

 

https://forums.unraid.net/topic/9277-how-to-report-a-defect-and-capture-syslog-and-smart-reports/

 

 

[trurl]

I haven't used this version in a very long time, and I didn't really do much (if any) support back then.

 

My guess is it is disk3 that is actually disabled, and disk2 is only missing. So it is really insisting that you put the original disk2 back so you can rebuild disk3.

 

That is my guess, maybe someone else will have another idea. Maybe we will know more after you get us the information I mentioned.

 

Don't do anything without further advice.

[AdamEdmonds]

Attaching syslog.  I'm not able to ssh to get smart reports. I will keep working on this.

syslog.zip

[AdamEdmonds]

Uploading disk3 Smart test information.  

disk3_Smart_Test.zip

[trurl]

That syslog has timestamps from February 17 (don't know what year).

 

Is your server clock correct?

 

Also, that syslog doesn't show anything about when the trouble began. Do you have earlier ones in the /var/log folder, such as syslog.1, syslog.2?

[JorgeB]

Disk3 is failing, but like mentioned if you can get disk2 back it will allow you to replace disk3, try connecting disk2 using different cables, swap with disk3 for example, both power and SATA cables.

[AdamEdmonds]

Thanks everyone for the suggestions.  I can say without question disk2 is completely dead.  There are loud mechanical noised during boot and the computer doesn't recognize the unit is plugged in regardless of cables or SATA port that is connected.  I will work on getting more log files.  

 

Edited for clarity and to remove typos.

Edited December 9, 2019 by AdamEdmonds

[JorgeB]

If disk2 is really dead you can't rebuild disk3, you can still recover the data from disks 1 and 4 by doing a new config, and you can use ddrescue with the old disk3 to recover as much data as possible, it won't recover any data written after the disk was disabled.

 

 

[AdamEdmonds]

Does anyone have any opinions/thoughts are sending disk2 out for recovery.  Would I be able to add the disk2 replacement into the array with hopes of rebuilding disk3?

[JorgeB]

2 minutes ago, AdamEdmonds said:

Would I be able to add the disk2 replacement into the array with hopes of rebuilding disk3?

Only if the returned disk is a bit by bit copy of the original, like when dd is used for example, if data is recovered by a some file recovery util and then written to another disk it won't work.

[AdamEdmonds]

Following this logic, if I could use ddrescue on disk3 I could potentially add it to the array and rebuild disk2, maybe?

[JorgeB]

1 minute ago, AdamEdmonds said:

if I could use ddrescue on disk3 I could potentially add it to the array and rebuild disk2, maybe?

Yes, but there will likely be some read errors, and that would result in a corrupt rebuild, how corrupt depends on the number of errors, but could be better than nothing.