OK here is the quick history.

 

1. I got a new SSD tonight and opened the case to install the new drive.  
2. Upon booting back up, one of my cache pool drives showed "not installed."  One of the two SSD's in my cache pool does not have a locking SATA connector.
3. I opened the case back up, checked the connections (felt ok) and rebooted.  
4. This time everything looked normal but within 5 minutes I started getting continuous stream red/write errors showing up on the monitor connected to unRAID direct. They all said BTRFS error and were some form of read/write.  Everything locked up, I couldn't get the machine to respond at all and finally had to do a hard shutdown.  
5. This time I checked the cables to the MB and the cache drives.  Again all felt fine but after rebooting it all looks fine, no read/write errors showing up.
6. EXCEPT Plexpy would not re-start, all my other dockers started fine.  When you tried to install Plexpy it said there was no such container.
7. I reinstalled Plexpy from the community apps page, and now everything really does look fine.  Even the Plexpy history is in place.
8. I canceled the parity check that tried to run after re-boot but will probably turn it back on tonight when I go to sleep, just to be safe.  I don't think the error actually had anything to do with the array itself so not overly concerned there.

 

So my question is, do I need to do anything else to confirm the integrity of the cache pool?   There isn't really a parity check to run (that I know of).  I'm most concerned about my Plex database, which again seems fine but no idea what the system did during those read errors.  I couldn't pull diagnostics before shutting down because it was frozen for 5 minutes or so completely unresponsive.

 

Thanks in advance.

 

Scrub your cache pool.

 

 

15 hours ago, johnnie.black said:

Scrub your cache pool.

 

 

 

Sorry I haven't been able to get to this all day.   What do you mean by scrub the pool?

Should I go back down to one device with my original device in the slot 1 position, start / stop the array, then go back to a cache pool and re-assign the 2nd device?

2 minutes ago, TODDLT said:

What do you mean by scrub the pool?

 

On the main page click on cache, on the cache page click on the scrub button.

5 minutes ago, johnnie.black said:

 

On the main page click on cache, on the cache page click on the scrub button.

 

Thanks, I never even knew that was there.

 

OUTPUT:   scrub started at Wed Aug 30 18:44:29 2017 and finished after 00:00:59 total bytes scrubbed: 20.12GiB with 0 errors

 

So that seems to be good news.  However, See the attached screen shot.  I have used 20 GB of space and have a pool of two 60 GB drives.  It used to show 60GB of space because the pool is mirrored.  However, it now shows 120GB of space.  Have i lost the mirrored drives?

 

Also, there is a "check file system status" as well that requires maintenance mode.  What is that?  (for future reference if not now).

 

 

 

I also see this, and it doesn't look right either.  

 

Quote

 

Label: none uuid: 0100252a-141c-446e-88ad-a90f2dba044d

Total devices 2 FS bytes used 20.08GiB

devid 1 size 55.90GiB used 12.03GiB path /dev/sdh1

devid 2 size 55.90GiB used 21.00GiB path /dev/sdi1

 

 

and this

 

Quote

 

Data, single: total=31.00GiB, used=19.79GiB
System, single: total=32.00MiB, used=16.00KiB
Metadata, single: total=2.00GiB, used=306.48MiB
GlobalReserve, single: total=41.38MiB, used=0.00B

btrfs balance status:

No balance found on '/mnt/cache'

 

What happened to the mirror?

 

 

Edited August 31, 20178 yr by TODDLT

I did try "balance and that just equally loaded both drives, but it still looks like an extended pool, not a mirrored pool.

Also why would on place show 15 GB used on each of two 60 GB drives and another show 20 of 120 used?

 

 

Label: none uuid: 0100252a-141c-446e-88ad-a90f2dba044d

Total devices 2 FS bytes used 20.12GiB

devid 1 size 55.90GiB used 15.03GiB path /dev/sdh1

devid 2 size 55.90GiB used 16.00GiB path /dev/sdi1

 

 

 

Pool is currently using single profile, not mirror, but you can convert to raid1:

 

 

9 hours ago, johnnie.black said:

Pool is currently using single profile, not mirror, but you can convert to raid1:

 

 

 

 

Thanks I did several searches last night looking for some instructions on cache modes and looked in the "official wiki" but as is obvious, I looked everywhere but where I needed to.  I should have known to look in the FAQ.

 

To be clear, I'm going to enter:  

-dconvert=raid1 -mconvert=raid1

in the "balance window" and then click "balance"

 

Do I need to make a copy of my appdata before executing this command?  Does it preserve everything in making the conversion?

 

As always thanks for help.

9 minutes ago, TODDLT said:

Do I need to make a copy of my appdata before executing this command?  Does it preserve everything in making the conversion?

 

All data is preserved, unless something goes wrong.

Done, thanks.  Closed issue.