[6.9.2] SMB transfer stops when pool is full and mover runs slow

[tronba]

Hi,
I just reinstalled my Unraid server, fresh copy of newest unraid version on the USB stick and formated all connected drives.
(5x4TB + 2x240 Kingston SSD.)
Everything is peachy until I try to copy files onto the new share i made (Using SMB).
Files start copying into the pool drives (speed around 80MBps so a bit low for a gigabit). I leave and when i come back the pool is full and the copy has stalled seemingly because of no free space in pool. (so it did not switch over to directly copying files to the array.

The pool also didn't start mover, when i did it the mover is only able to copy around 40MBps (very slow compared to before upgrade to new unraid version)
Anyone have a suggestion about what this can be?

nasse-diagnostics-20210701-1810.zip

Edited July 1, 2021 by tronba
Added detaile

[trurl]

3 minutes ago, tronba said:

the pool is full and the copy has stalled because of no free space. (so it did not switch over to directly copying files to the array.

You need to set Minimum Free for cache

[JorgeB]

4 minutes ago, tronba said:

I leave and when i come back the pool is full and the copy has stalled because of no free space.

That will happen if the minimum free space for the share is not set to more than the largest file you copy to that share, usually recommendation is to set to twice that.

 

5 minutes ago, tronba said:

the mover is only able to copy around 40MBps

That's about right for default writing mode, see turbo write.

[trurl]

Looks like you have corrupted cache by overfilling

Jul  1 08:51:27 Tower kernel: BTRFS warning (device sdb1): csum failed root 5 ino 264 off 1637920768 csum 0xc05895db expected csum 0xd8ca49a2 mirror 1
Jul  1 08:51:27 Tower kernel: BTRFS error (device sdb1): bdev /dev/sdb1 errs: wr 0, rd 0, flush 0, corrupt 1, gen 0
Jul  1 08:51:27 Tower kernel: BTRFS warning (device sdb1): csum failed root 5 ino 264 off 1637920768 csum 0xc05895db expected csum 0xd8ca49a2 mirror 2
Jul  1 08:51:27 Tower kernel: BTRFS error (device sdb1): bdev /dev/sdc1 errs: wr 0, rd 0, flush 0, corrupt 1, gen 0
Jul  1 08:51:27 Tower kernel: BTRFS warning (device sdb1): csum failed root 5 ino 264 off 1637920768 csum 0xc05895db expected csum 0xd8ca49a2 mirror 1
Jul  1 08:51:27 Tower kernel: BTRFS error (device sdb1): bdev /dev/sdb1 errs: wr 0, rd 0, flush 0, corrupt 2, gen 0
Jul  1 08:51:27 Tower shfs: copy_file: /mnt/cache/Arkiv/Backup4/En måned i naturen/1 mnd i naturen/dag 1/569_0902_01.mxf /mnt/disk1/Arkiv/Backup4/En måned i naturen/1 mnd i naturen/dag 1/569_0902_01.mxf (5) Input/output error

and so on

[trurl]

9 minutes ago, tronba said:

pool also didn't start mover

Mover is by schedule, default daily in the middle of the night. There is a plugin to run mover according to other criteria.

[trurl]

5 minutes ago, trurl said:

You need to set Minimum Free for cache

Click on first cache disk to get to this setting.

[tronba]

Thank you all for the fast feedback.
I will increse the minimum free space and check up on turbo write.

With kind regards.
T

[JorgeB]

17 minutes ago, trurl said:

Jul 1 08:51:27 Tower kernel: BTRFS warning (device sdb1): csum failed root 5 ino 264 off 1637920768 csum 0xc05895db expected csum 0xd8ca49a2 mirror 1

These are data corruption errors, you should run memtest.