I've been fighting this issue since my upgrade from 6.9.2 to 6.11.1.  My Ventura OSX stoped backing up to my TimeMachine share.  I tried absolutely EVERYTHING on this earth available here on this forum or reddit,  but no luck.  Even tried the TimeMachine docker and still no luck.  It works for a while and then it stops working.  I even redone may backups from scratch several times, even moved my TimeMachine to a second UnRaid, but  I have with the same disappointing results.  All other Mac versions work fine except the Ventura OSX. I am so pissed off with the amount of time I've spent on this. Absolutely ridiculous  

 

I am officially giving up on UnRaid to backup my 8 Macs and moving back to my Mac Mini with a thunderbolt enclosure and raid.  s just reliable and it works.

I have the same problems with TimeMachine.... sometimes it works, sometimes it doesn't.

 

Recently I discovered that many of my working backups were done in the night around 3:30, when I was not using the computer at all. But my Unraid server does have a cronjob at 3:30, which does a rsync backup of another machine. 

 

So today I made an experiment:

Last working TimeMachine backup on my Mac was 3 days ago, manually starting the backup did not work. I tried that multiple times in a row.

Then I run the rsync-script on my Unraid server and after that run TimeMachine on my Mac again. And magically it worked!

 

I repeated the whole thing twice, TimeMachine seems only to connect to the backup volume if that rsync script is working in background. But when the script was running, TimeMachine worked every time. My Unraid machine is quite small (2 HDDs+1 Cache-SSD for now) and not accessed most of the time. Could it be that something is idle and therefore TimeMachine runs in a timeout and cannot connect?

 

@limetech I will attach logs and my configuration here, just in case it helps...

 

SMB Settings

Enable SMB: Yes – Hide dot files: No – SMB Multi Channel: Yes – Mac interop.: Yes – NetBIOS: No – WSD: Yes

 

/boot/config/smb-fruit.conf

vfs objects = acl_xattr fruit streams_xattr
fruit:encoding = native
fruit:metadata = stream
fruit:posix_rename = yes

 

And here is the rsync-script, its basically only cleaning up some older backups and then calling rsync, so I guess it is the workload on the server which makes the difference:

#!/bin/sh
DAY=$(date +%A)
if [ -e /mnt/user/server/$DAY ] ; then
  rm -fr /mnt/user/server/$DAY
fi
rsync -qaHAXS --delete --inplace --backup --backup-dir=/mnt/user/server/$DAY [email protected]:/home [...] /mnt/user/server/current/

 

Edit: I experimented with different approaches to put load on the server.

 

TimeMachine backup works, if:

- rsync script is running (see above)

- I stress test my cpu with this command: 

for i in $(seq $(getconf _NPROCESSORS_ONLN)); do yes > /dev/null & done

 

TimeMachine backup does not work, if:

- Unraid server is not in use

- I copy large files from Unraid to my Mac

- I copy large files from my Mac to Unraid

nas-diagnostics-20230122-1201.zip

Edited January 22, 20233 yr by Simon Edelmann
Adding more information...