Updated to 6.12, can no longer access shares from Windows or Linux machines

[Lignumaqua]

Updated to Unraid 6.12 yesterday. All seemed to be running correctly, however I can no longer access shares from Windows or Linux machines. Tried 4 different Windows computers (3 virtual on 2 separate servers and one bare metal Windows machine), all running Windows 10 and 11 and none of them can access Unraid shares. Same with Linux. The Unraid machine is visible in the Windows Network list, but attempting to access it results in a 'Windows cannot access \\TOWER, Error code 0x80070035 The network path was not found" message. Linux gives me a 'Failed to receive shares list' message.

Also tried creating a new connection from Windows using the direct ip address rather than the name, and that doesn't work either.

 

Oddly, although Windows and Linux don't work, I CAN access shares from a Mac.

 

Is there a change in how shares work in 6.12 that would affect connection from Windows and Linux machines?

Diagnostics attached. Thank you.

tower-diagnostics-20230616-1349.zip

[Lignumaqua]

I tried the brute force approach, and restarted the smb service. That seems to have fixed it! 🙂 No ideas why smb fell over, but fall over it did. Strange.

 

So, I'm back running again, no need for urgent help now unless you can help diagnose why smb crashed in the first place. Thanks!

[JorgeB]

Jun 16 01:19:23 Tower kernel: macvlan_broadcast+0x10a/0x150 [macvlan]
Jun 16 01:19:23 Tower kernel: ? _raw_spin_unlock+0x14/0x29
Jun 16 01:19:23 Tower kernel: macvlan_process_broadcast+0xbc/0x12f [macvlan]

 

This is unrelated to your issue but good idea to change from macvlan to ipvln, or these call traces may end up crashing the server.

[bonienl]

You have a routing loop.

In your network configuration make sure the gateway for br0 has a LOWER metric value then the gateway for br1.

 

[Lignumaqua]

Thank you both!

[Lignumaqua]

5 hours ago, JorgeB said:

Jun 16 01:19:23 Tower kernel: macvlan_broadcast+0x10a/0x150 [macvlan]
Jun 16 01:19:23 Tower kernel: ? _raw_spin_unlock+0x14/0x29
Jun 16 01:19:23 Tower kernel: macvlan_process_broadcast+0xbc/0x12f [macvlan]

 

This is unrelated to your issue but good idea to change from macvlan to ipvln, or these call traces may end up crashing the server.

FYI - I shut down Docker service, made the change, and re-enabled Docker but the Docker service didn't start. I had to use the console to force docker start. As an experiment I then tried to start and drop Docker from the GUI again without making any changes. The same thing happened. Docker seems to start fine from command line, but not from GUI. Don't know if this is a bug or just my system? Now running with ipvlan as recommended. Thank you.

 

 

3 hours ago, bonienl said:

You have a routing loop.

In your network configuration make sure the gateway for br0 has a LOWER metric value then the gateway for br1.

 

 

I believe it already is. br0 has a metric of 1011 and br1 is 1012 in the 'Routing Table' section of Network settings.