Last week I upgraded my cache settings from a single cache drive to a dual cache drive (included snapshot). That view makes it look like the second drive is idle and sundown. Is that just showing misleading information? I don't know how to verify I'm truly operating as RAID 1. These are two 1TB drive and df shows it as a 1TB filesystem.

 

root@laffy:/mnt/cache# df -h .
Filesystem      Size  Used Avail Use% Mounted on
/dev/sdb1       932G   47G  884G   5% /mnt/cache
 

Please post your diagnostics to confirm but it does look like it's not being used: Tools -> Diagnostics

Attached. Thanks for taking a look at this.

laffy-diagnostics-20180901-0944.zip

Yes, cache2 is not being used, this should fix it but backup any important data on cache before starting just in case it goes wrong:

 

If you rebooted since diags check cache2 is still sdg, then stop the array and on the console type:

 

wipefs -a /dev/sdg1

then

wipefs -a /dev/sdg

 

Click on the GUI main page to refresh it, wiped device icon should turn blue
Start array, balance should begin

Thanks! I'll review those steps later this afternoon. Will have to move some Docker stuff off of there to be safe and will attempt that. Will report back on findings.

 

How should the main screen look when redundancy is working?

1 hour ago, snowboardjoe said:

 

How should the main screen look when redundancy is working?

 

Very similar to your original screen grab but with more meaningful read/write counts.

 

 

 

Shutdown Docker.

Migrated all data off of cache drive.

Ran wipefs commands on sdg1 and sdg successfully.

Moved data back to cache drive.

Restarted Docker.

Reset disc statistics on Main screen.

 

Still no write operations to the second cache drive.

11 hours ago, snowboardjoe said:

Ran wipefs commands on sdg1 and sdg successfully.

Did you refresh the page before starting, did the cache2 icon turn blue before starting the array?

If yes and it still didn't work try unassigning cache2, start the array, stop the array, reassign cache2 and start again.

 

When done you check if it's working by generating new diags and looking at system/btrfs-usage.txt, this is from previous ones:

 

Overall:
    Device size:         931.51GiB
    Device allocated:          50.02GiB
    Device unallocated:         881.49GiB
    Device missing:             0.00B
    Used:              46.37GiB
    Free (estimated):         883.23GiB    (min: 442.48GiB)
    Data ratio:                  1.00
    Metadata ratio:              2.00
    Global reserve:          33.83MiB    (used: 0.00B)

Data,single: Size:48.01GiB, Used:46.27GiB
   /dev/sdb1      48.01GiB

Metadata,DUP: Size:1.00GiB, Used:51.56MiB
   /dev/sdb1       2.00GiB

System,DUP: Size:8.00MiB, Used:16.00KiB
   /dev/sdb1      16.00MiB

Unallocated:
   /dev/sdb1     881.49GiB

Only sbd is part of the pool, when it's correct both sdb and sdg will appear.

Oh? I was never told to stop the array. I did not do that. I will try it again later this morning using the procedure you described.

10 minutes ago, snowboardjoe said:

I was never told to stop the array.

 

Actually, you were:

 

21 hours ago, johnnie.black said:

If you rebooted since diags check cache2 is still sdg, then stop the array and on the console type:

 

Sorry, that was not clear. Your message implied that I needed to shutdown if the device name had changed (and it hadn't). There is a lot about the cache process that is complicated and not documented. Wish there was a better way to evaluate the config with a command before firing up the array.

 

I'll go through the process again and see how that goes later this morning or possibly tomorrow. Which process should I follow? Follow your original steps and wipe it again while the array is stopped? 

Try the first option first, more likely to work, don't forget to refresh the GUI after the wipe, cache2 icon should turn blue (new disk)

• Stopped Docker.
• Migrated all data off of cache and verified empty.
• Shutdown array.
• Ran wipefs commands as stated above:
  • /dev/sdg1 was not found (not surprised there)
  • No output from wiping /dev/sdg (returned 0 exit code)
• Refreshed main screen.
• Cache 2 still has a green dot (never went blue).

Suggestions on next steps? Remove, format and re-add?

Try the other option, unassign and reassign.

Unassigned and reassigned. Still green.

Unassigned, formatted and reassigned. Still green.

Something weird going on, but best bet is to wipe both cache devices and start over.

Got it! While array was still stopped:

• Removed second cache drive.
• Set cache slots to 1.
• Set cache slots back to 2.
• Assigned second cache drive (blue icon!)
• Started array.

I now see R/W activity for both devices.

👍