Hi All,

 

7.0.0 beta 2..

So stupid me, I was using the cacheshare as a share and storing some files I accessed frequently since it's a SSD.

It started to get full, so I removed all those files. Not sure how to easily see how much data is ACTUALLY on there now but it's a 2 TB SSD, if I access the share and do a DIR in dos, it shows 2 GB free (should be MUCH MUCH more)... 

Syslog is full of

Oct  4 08:35:30 Tower shfs: assign_disk: /appdata/kodi-headless/userdata/Database/Addons33.db-journal (28) No space left on device
Oct  4 08:35:31 Tower shfs: share cache full
every other second - same file name.

I ran mover a few times since - when I ran it from the web interface, as soon as I clicked Move, and the page refreshed, it went back to not running (ie the icon was available to run again.
I tried mover from terminal - it finished in about half a second.

 

I now tried it again, and it appears to be running (web gui says it's running)

I haven't seen any entries in the syslog regarding mover.

 

- Update on the DIR command - when I started typing this post, my cacheshare showed 2,675,277,824 bytes free. A few minutes later, 1,886,445,568 free.

 

Any suggestions on what to look for?

Thanks.

Enable the mover logging, run the the mover, post the diagnostics and he share name(s) you want moved.

Here's the logs, thankstower-diagnostics-20241004-0956.zip

 

I enabled mover logging, I assume that's included in the diags? I didn't see anything extra added to syslog.

 

Thanks

Share name is cacheshare but to clarify - it SHOULD be a cache, but now I removed a TON of data from it so it should be nowhere near full.

Edited October 4, 20241 yr by phirephoto

Cacheshare is set to cache=only so nothing will be moved, see here:

 

https://docs.unraid.net/unraid-os/release-notes/6.12.0/#share-storage-conceptual-change

 

 

So I was using it as a share but removed the majority of data on it. But cache array is STILL full. 2 TB of data I can't see....

45 minutes ago, phirephoto said:

So I was using it as a share but removed the majority of data on it

Cache is 100% full, so you didn't remove any data, or it filed up again.

So.... My shares are all set to cache -> Array.

my share 'cacheshare' is a share on the cache drive.

If I moved data from /cacheshare to a regular drive,  would it move it directly to the array drive or would it go through the cache first?

My GUESS is I was moving data from the cache to the cache (which would then be filled up the cache even more), and something got 'stuck'... just not sure how to track down where or what is stuck - if that makes sense?

Ok it looks to me like data IS stuck on the share... Before I delete it, want to get a input so I don't screw anything up.

My computers - I see the cacheshare and I had a folder called ROMS.

That folder no longer 'exists' as far as my computers see - ie \\IP\CACHESHARE\roms .

BUT /mnt/cache/DRIVE/roms - I see the files.

DRIVE is my NAS share (cache -> array). I moved the files from \\IP\cacheshare\roms to \\ip\drive\roms.  But in /mnt/cache/drive/roms is 650GB. Can I mv these files manually? 

this definitely sounds like a mover issues. IN v7 beta 2 I've found that mover refuses nor doesn't complete nor run sometimes. (Especialy when moving data or the array / cache to move to zfs only...

also i would have you take a look at plugin mover Turner to set other settings to test

I had to open unraid web terminal and manual move the files with MC. beta2 has something broken when calling mover

*Found this 2 days ago and trying to dig into how and why mover failed. there may already be a bug report...

Edited October 4, 20241 yr by bmartino1

we would need more data, a diag file, picture of the shares tab and main tab to help diagnose further. we need to know what your unraid settign are to figure it out...
 

potential fix is to add something to the go file to restor mover:

 

Thanks to the people in this thread: https://forums.unraid.net/topic/70783-plugin-mover-tuning/page/59/

Edit your go file at: /boot/config/go and add:

rm /usr/local/sbin/move

ln -s /usr/libexec/unraid/move /usr/local/sbin/move

14 hours ago, phirephoto said:

If I moved data from /cacheshare to a regular drive,  would it move it directly to the array drive or would it go through the cache first?

How were you moving the data? You can use the mover if you set the share correctly.

I tried 

rm /usr/local/sbin/move

ln -s /usr/libexec/unraid/move /usr/local/sbin/move

in /boot/config/go and even ran the command manually, but my mover appears broken. My cache drive is 100% full and I cannot write to any shares that are cache -> array ....

Do you have the mover tuning plugin installed?

Yes.. I even just downgraded to beta 2 and ran a manual move, but nothing's moving.. 65 kb free  on the cache drive  

Reboot to clear the logs, enable mover logging, run the mover and post the diags, together with the share name(s) you expect ot be moved.

tower-diagnostics-20241007-1036.zipBefore I reboot.. do you see what's taking a literal 45 minute startup time? Any time I reboot, it takes 30-45 minutes for the unraid to come back online. Rebooted yesterday at 2 PM, it was 2:45 before it came back online?

Hoping to maybe fix that before I do this reboot again. THanks

Devices mounted and services started a couple of minutes after the boot start, so if there's a problem it may be external.

Here's my logs.. all shares are set as cache -> array 

 

DISK2

Ddomains

DRIVE

DropBox

isos

system

timemachine

tower-diagnostics-20241007-1317.zip

So removing the mover plugin seems to have resolved it...

10 hours ago, phirephoto said:

So removing the mover plugin seems to have resolved it..

Yes, that is known to cause issues, and why I asked, but misread your reply, thought you had already removed it.