Thanks for your input 
I think I switch 1 node drive to ext4 and compare the behavior.
This will take a while but hopefully its better to win more data races.
Is there any disadvantage by using a single rsync command for the whole storj folder?
longer downtime than 2 waves of rsync.
1 Like
If you would do that without stopping the node, you may do not catch changes happened during rsync, and without a final rsync with --delete flag when the node is stopped you likely will end up with corrupted databases (because temp files for opened databases have not been removed).
Sorry, I have expressed myself incorrectly. So of course I want to run Rsync 2x in operation and then run --delete with stopped node. But instead of running each folder individually I want to copy the entire Storj folder.
That should not be a problem, right?
Yes, of course you need to copy the entire folder, not one by one. Out of curiosity - why did you think it should be copied one by one?
The only separate could be your identity folder, unless you moved it to the data location too. In this case you may rsync everything at once.
1 Like
I just wanted to be safe 
First rsync run is sucessfully done but the second stops after 5 seconds.
command
rsync -aP /mnt/disks/14TB_WD_Storj_1/Storj/ /mnt/disks/12TB_WD_Storj/Storj_14TB_WD_Storj_1/
output
4,120,032 100% 22.58MB/s 0:00:00 (xfr#7, ir-chk=1072/1705)
data/storage/storage_usage.db-shm
32,768 100% 183.91kB/s 0:00:00 (xfr#8, ir-chk=1039/1705)
data/storage/storage_usage.db-wal
173,072 100% 965.80kB/s 0:00:00 (xfr#9, ir-chk=1038/1705)
After data/storage/storage_usage.db-wal it does not get any further. Stopped the node but same result.
Perhaps you do not have permissions. You need to run this command under owner.
Please check owners:
ls -l /mnt/disks/14TB_WD_Storj_1/Storj/
1 Like
output
root@unraid:~# ls -l /mnt/disks/14TB_WD_Storj_1/Storj/
total 0
drwxrwxrwx 3 root root 30 Jun 10 2020 config/
drwxrwxrwx 4 root root 124 Oct 19 06:22 data/
I run every command as root.
edit:
I let the tmux command prompt open over night and now it works again, but its really slow seems to take a lot more time. For my understanding it should be a lot faster, because the most data is already copied.
Unfortunatelly it need to check every copied file to make sure that the file is properly copied. When you have a lot of small files (our case), it takes most of the time rather than copying itself.
1 Like
Thank you good to know. So it should now take as long as the first run?
1 Like
It could be shorter, but not predictable. You may reduce the allocation below the usage for this node to do not accept any new load, then it should take much less time.
1 Like
First Transfer is done, drive formatted to ext4. The copy speed from a XFS to ext4 drive is more than 2x as fast! Thank you for your time and help!
1 Like