SSD cache pool only utilises 1 disk?

[robertoal]

See attachment for details.

Is what is shown in the attachment normal?

I would suspect, because there are 2 drives, that raid 1 would be used; and so both drives should have identical writes/reads?

 

 

 

[JorgeB]

Looks like only one is being used, post your diagnostics.

[McKentin]

Change you cache pool raid settings. Easy fix

[robertoal]

Maybe stupid.. but where can I find those settings?

 

Attached dianostics!

 

Thanks guys

unraiddata-diagnostics-20170808-1624.zip

Edited August 8, 2017 by robertoal

[JorgeB]

You're using an older version that doesn't include btrfs pool info on the diagnostics, update to v6.3.5 and post new diags.

[robertoal]

Thanks for the tip John, I attached the new diagnostics file!

unraiddata-diagnostics-20170809-1101.zip

[JorgeB]

The diagnostics confirm only one device (cache2) is being used, this should fix it:

 

-Backup your cache in case something goes wrong

-Stop the array

-if you rebooted since check that cache1 is still sdf and wipe it using:

 

blkdiscard /dev/sdf

-on the WebGUI click on "Main", cache1 SSD icon should turn blue

-start the array, a balance will begin and when all cache pool read/writes stop the pool should now be using both devices.

 

You can post new diags to confirm if unsure.

[robertoal]

Johhny thanks for your help again!!

Check diagnostics i attached, looks like its working now!

 

+1 Rep!

unraiddata-diagnostics-20170816-1430.zip

[JorgeB]

Both are being used now but it's not balanced correctly, on the main page click on "cache", then on "balance status" paste in the options:

 

-dconvert=raid1 -mconvert=raid1

 

Then click on balance.

[robertoal]

It is balancing now. thanks again!

I am planning on getting in two more 160gb intel ssd's, especially for VM's the room I have now is not enough. Which options should I use then?

[JorgeB]

With 4 devices you can use raid10, you get redundancy, capacity and better performance.

[robertoal]

Good to know. Much obliged !

[robertoal]

Johnie! In need of your help again.. 

 

Added 2 more 160gb intel ssd's but balancing does not work anymore, ends with: 

No balance found on '/mnt/cache'

Log shows:

Aug 17 11:56:53 UnraidData kernel: BTRFS warning (device sdi1): csum failed ino 354 off 623747072 csum 1316965508 expected csum 4252368761
Aug 17 11:56:53 UnraidData kernel: BTRFS warning (device sdi1): csum failed ino 354 off 623747072 csum 1316965508 expected csum 4252368761
Aug 17 11:56:53 UnraidData kernel: BTRFS warning (device sdi1): csum failed ino 354 off 623747072 csum 1316965508 expected csum 4252368761
Aug 17 11:56:53 UnraidData kernel: BTRFS warning (device sdi1): csum failed ino 354 off 623747072 csum 1316965508 expected csum 4252368761
Aug 17 11:56:53 UnraidData kernel: BTRFS warning (device sdi1): csum failed ino 354 off 623747072 csum 1316965508 expected csum 4252368761
Aug 17 11:56:53 UnraidData kernel: BTRFS warning (device sdi1): csum failed ino 354 off 623747072 csum 1316965508 expected csum 4252368761
Aug 17 11:56:53 UnraidData kernel: BTRFS warning (device sdi1): csum failed ino 354 off 623747072 csum 1316965508 expected csum 4252368761

 

Tried both: 
-dconvert=raid1 -mconvert=raid1

 

and

-dconvert=raid10 -mconvert=raid10

 

Attached diags again!

 

edit: Tried scrubbing.. but ended with 3 non recoverable errors. 

 

 

unraiddata-diagnostics-20170817-1203 (1).zip

Edited August 17, 2017 by robertoal

[robertoal]

Ok. Somehow it was unable to find sdh.. on which the data was located. Thank god I make backups :-)

Now I just removed al partitions and let it recreate them. 

Think it works now although it still says 'no balance found'

Is this allright?

[JorgeB]

40 minutes ago, robertoal said:

Think it works now although it still says 'no balance found'

Is this allright?

 

Yes, it only shows different info during a balance.