Error state on drive

[highgear]

Hello all,

 

I just discovered that one of my drives is disabled and there are 2 errors on the drive. I suspect it's a cable issue but I'm not 100%. Everything is properly seated. I do have some new cables that I can swap out as well as cold pre cleared disks if needed. I was considering replacing the cable and rebuilding the drive onto itself. Any thoughts here? I've attached my diagnostics.

 

Thanks in advance I appreciate it!

 

tower-diagnostics-20230213-0843.zip

[trurl]

Just to make sure we aren't missing anything, why do you not have anything assigned as disks 1, 2?

 

[JorgeB]

Disk dropped offline so there's no SMART, check/replace cables and post new diags.

[highgear]

1 minute ago, trurl said:

Just to make sure we aren't missing anything, why do you not have anything assigned as disks 1, 2?

 

Sure. I just recently shrunk my array and removed a few drives. I used the newer process and used the user script to clear the drives and then removed from the array. There was a parity check completed with 0 errors after. This was a few weeks ago and those drives are under unassigned devices.

[trurl]

Disabled/emulated disk3 is mounted, so that's good.

 

But no SMART report for disk3, looks like it has disconnected.

2 minutes ago, JorgeB said:

Disk dropped offline so there's no SMART, check/replace cables and post new diags.

 

[highgear]

2 minutes ago, JorgeB said:

Disk dropped offline so there's no SMART, check/replace cables and post new diags.

Ok will do shortly. Thanks.

[highgear]

1 hour ago, highgear said:

Ok will do shortly. Thanks.

Replaced the cable on the disabled drive. It's now showing as connected and the SMART is included in the new diags attached. Thanks for taking a look.

 

@trurl

@JorgeB

 

tower-diagnostics-20230213-1028.zip

Edited February 13, 2023 by highgear

[JorgeB]

SMART looks good, assuming the emulated disk contents look correct you can rebuild on top.

 

https://wiki.unraid.net/Manual/Storage_Management#Rebuilding_a_drive_onto_itself

[trurl]

23 minutes ago, JorgeB said:

assuming the emulated disk contents look correct

Can't tell from these latest diagnostics since the array wasn't started. Previous diagnostics had emulated disk3 mostly full. Some other disks mostly full too.

 

So after you rebuild, you need to plan for expansion.

[highgear]

39 minutes ago, JorgeB said:

SMART looks good, assuming the emulated disk contents look correct you can rebuild on top.

 

https://wiki.unraid.net/Manual/Storage_Management#Rebuilding_a_drive_onto_itself

 

Ok sounds good, thanks! Rebuilding onto the drive now. I'm currently only getting speeds of around 50 MB/sec on the rebuild though. Docker is disabled and there shouldn't be any other reads/writes happening that I'm aware of. For parity checks I normally get 100MB+ speeds. Any thoughts?

 

 

10 minutes ago, trurl said:

Can't tell from these latest diagnostics since the array wasn't started. Previous diagnostics had emulated disk3 mostly full. Some other disks mostly full too.

 

So after you rebuild, you need to plan for expansion.

 

Thanks yeah I have plenty of extra space currently but my allocation settings are a bit wonky and always fill my drives up more than I'd like. What's a reasonable amount of minimum free space on each drive you like to see?

[trurl]

2 minutes ago, highgear said:

Any thoughts?

Post new diagnostics 

[highgear]

2 minutes ago, trurl said:

Post new diagnostics 

Thanks. I'm now seeing it cycle between 30-90ish MB/sec.

tower-diagnostics-20230213-1146.zip

[itimpi]

23 minutes ago, highgear said:

Thanks yeah I have plenty of extra space currently but my allocation settings are a bit wonky and always fill my drives up more than I'd like. What's a reasonable amount of minimum free space on each drive you like to see?

The general rule of thumb is something like twice the size of the largest file you normally expect to write to the share.   Another thing to think about is that if for any reason you get file system corruption on the drive it is a good idea to have something like 10-20 GB free for the check/repair process to use. 

[JorgeB]

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
 1271 root      20   0    3804   2360   2080 R 100.0   0.0   0:00.77 find

 

This is using a lot of CPU and likely looking for something in the array.

[trurl]

50 minutes ago, JorgeB said:

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
 1271 root      20   0    3804   2360   2080 R 100.0   0.0   0:00.77 find

 

This is using a lot of CPU and likely looking for something in the array.

Maybe this?

dynamix.cache.dirs.plg - 2023.02.04  (Up to date)

 

[highgear]

5 hours ago, itimpi said:

The general rule of thumb is something like twice the size of the largest file you normally expect to write to the share.   Another thing to think about is that if for any reason you get file system corruption on the drive it is a good idea to have something like 10-20 GB free for the check/repair process to use. 

Ok thanks that makes sense.

 

4 hours ago, trurl said:

Maybe this?

dynamix.cache.dirs.plg - 2023.02.04  (Up to date)

 

Ok cool I’ll stop that in the future. Chugging along at 160MB now so looking good. Thanks for the help everyone.