Hi everyone.

I have problem with my virtual mashine. 2 months ago i switched from my phisical  server to virtual mashine. i got 1 TB phisical disk and i converted it to qcow2 dynamic disk. After this conversion qcow2 file was around 250GB.  I created virtual mashine in unraid and put my qcow2 file on new 2x nvme disks specialy bought for my VM's and and dockers and formated with btrfs (as far as i know this is only option for disk redundancy).  and all was working fast and ok. 

 

I'm  making backup of this file regulary once in the week. Procedure is simple i shut down VM. than use rsync -ah .....  command to copy file from my nvme disk to array. At  the beggining coppy speeds was around 166MB/s almoust constantly (all backup took like 25-30 min).

 

but after this i started to put some files inside VM and my qcow2 file grow up.  first 4 backups was rather fast but took me longer and longer every next week. i also expirienced more and more  not 166MB/s speeds but 3MB/s !!. Right now after 2 months of usage my qcow2 file have 278GB and my speeds are very slow now. mousty limited to magical 3MB/s and i dont know why. backup take up like 1:20 hours now.

average speeds dropped to 50MB/s.  i know my disks are good in array and nvme. 0 smart warrnings or errors. I wonder if this is a problem with dynamic qcow2 disk or with nvme formated  with btrfs.  Can any 1 help what is going on? should i use fixed sixe qcow2 file or try to move vm on xfs disk?

 

i moved appdata, domains, system, and, dockers folders also to thesame nvme cache pool what my qcow2 file is stored. Do i do something wrong here? 

 

my VM is windows and it makes alot of writes to swap file. maby i should balance or scrub nvme pool disk more 

those are the statistics of  nvme pool 

 

Data, RAID1: total=358.00GiB, used=334.84GiB System, RAID1: total=8.00MiB, used=80.00KiB Metadata, RAID1: total=1.00GiB, used=501.12MiB GlobalReserve, single: total=448.27MiB, used=0.00B

btrfs balance status:

No balance found on '/mnt/cache'

 

 Current usage ratio: 93.5 % --- No Balance required

 

UUID: f1ed6f74-2168-4a95-a1ed-a1a37537a84f no stats available

Total to scrub: 670.66GiB

Rate: 0.00B/s

Error summary: no errors found

 

 

 

unraid version is updated to 6.12.6 but this not solved the problem.

You can try rsync --inplace, it will only send the changed parts of the file, it may be faster, thought using snapshots and send and receive should be much faster, but for that you'd need a btrfs disk in the array.

 

12 hours ago, MoherPower said:

(as far as i know this is only option for disk redundancy)

You can also try zfs to see if there's any difference, and again backups would be more efficient if you then use snapshots with send/receive

thank you for answer. i will try zfs and get back in fiew days.