TimeMachine unavailable until SMB restart

TrueNAS General
, ,
1

I’m running truenas Scale 23.10.2 and my MAC client has always been able to use TrueNAS as a target for its TimeMachine backups.

Then, yesterday, I noticed that this client had not done any backups since April.
I tried to run TimeMachine manually but it said it couldn’t connect to my truenas server.

I then logged into Truenas and restarted SMB. Sure enough, after doing this the MAC client was able to connect to TrueNAS and was able to complete its TimeMachine backup.

Now, have any of you seen this behavior before?
I would like to troubleshoot this since I need to be able to trust that the backups are running properly without my intervention.

Help?!

2

Is this a one time issue or have you seen this before?

You might set up a job to restart SMB at a maintenance window.

Check the release notes on Scale versions to see if it is mentioned.

3

Actually, this has happened once before.
Back then (and i don’t remember how many months ago was this) it happened and i couldn’t get the MAC to connect to TrueNAS so i destroyed the dataset/share and created new ones.

Then, the MAC was able to start the TimeMachine backups. When i did that, somebody suggested that what may have worked back then was the fact that when you create a new share the SMB service is restarted.

That is why i started this just by restarting SMB to see if that would fix the problem, and sure enough it did.

Now, do you happen to know how to start troubleshooting this?
I mean, i cannot fix something if i don’t know what is the problem and what produced it.

4

I just know MacOS, SMB and problems are common topics. You can try searching the forums and see any of the other posts are helpful

5

Well, start by trying to reproduce on 24.04.2, Ie the current release of TrueNAS

I’ve had similar issues I think, and next time I’ll try the SMB cycling “fix”

6

@Stux , I’m in the process of upgrading. I just need to migrate one last app from the truecharts catalog to a custom app and then I’ll do the upgrade.

Hopefully the upgrade itself is going to fix this issue.

1 Like
7

Well, I have Updates.
While on 24.04 i had not seen any problems with TimeMachine.

I am now in 24.10.2 and the MAC user reported that their machine had not been able to find the TimeMachine target (i.e. TrueNAS server). I restarted SMB in trueNAS and the MAC user was able to find it right away and started doing its backup.

So, i would say that this seems to be a bug unless someone tells me a workaround to fix this?

8

Maybe try just restarting avahi. mDNS issues can present this way and may be environmental.

9

@awalkerix i think you are on to something.

I noticed that the MAC machine identified the TimeMachine target as the hostname of my truenas server but with “.local” as the domain.

This is weird because my local domain is “home.arpa” and is distributed by the DHCP server in my firewall (OPNsense).

Does the “TimeMachine” share have a different domain name setting than the main truenas server?

10

https://www.rfc-editor.org/rfc/rfc6762

MacOS by default uses mDNS for resolving time machine shares. You can manage things directly by using tmutil on the mac. Various network design issues can make mDNS unstable which manifests as the server “disappearing”. Some vendors work around this by basically restarting mDNS every few seconds.