This just started. I cannot figure out how to reproduce it. It acts like it just starts blocking all incoming connections. I can fix it every time by pinging anything from the unraid server. I have tried multiple cables and even a usb to ethernet adapter. There are other computers on the switch its connected to and all of those are working fine while the issue on unraid keeps happening.

diagnostics.zip

Jul 7 16:56:19 Tower kernel: r8169 0000:2a:00.0 eth0: Link is Down

Was this you disconnecting the cable?

6 hours ago, JorgeB said:

Jul 7 16:56:19 Tower kernel: r8169 0000:2a:00.0 eth0: Link is Down

Was this you disconnecting the cable?

No. The diags posted are from new boot to issue happening. I did nothing except start array.

Edited July 8, 2025Jul 8 by jmztaylor

That looks like your switch is dropping the link.

That suggests a NIC, cable, or switch issue.

6 minutes ago, JorgeB said:

That suggests a NIC, cable, or switch issue.

Already replaced/swapped all.

It happens with a brand new cable.

It happens with a usb network adapter

It happens when plugged into router instead of switch

Something interesting though. Any dockers set to Custom: br0 are still accessible.

Edited July 8, 2025Jul 8 by jmztaylor

Not sure what else could be, it's losing the physical link.

3 minutes ago, JorgeB said:

Not sure what else could be, it's losing the physical link.

But when Unraid dashboard/ssh and dockers are unavailable the Pi hole with network set to custom: br0 is working without issue.

It also seems no settings in unraid are setting after hitting save.

So I think I should rebuild my USB.

EDIT: I created new usb, copied over the disk assignments and pools, along with docker previous apps folders. Still having same issue.

Edited July 8, 2025Jul 8 by jmztaylor

1 hour ago, jmztaylor said:

But when Unraid dashboard/ssh and dockers are unavailable the Pi hole with network set to custom: br0 is working without issue.

Not sure how anything can work with the NIC link down, do the issues always coincide with link down messages, or sometimes are they not there?

16 minutes ago, JorgeB said:

Not sure how anything can work with the NIC link down, do the issues always coincide with link down messages, or sometimes are they not there?

Can you check disk 6 in the diags? It's been kicked off. Is there any chance a failing/failed disk might be killing stuff randomly?

Realtek nic by chance ? , maybe try the drivers from CA , search for Realtek

3 hours ago, Rhysee said:

Realtek nic by chance ? , maybe try the drivers from CA , search for Realtek

That didn't help.

So some new info.

Drive 6 is done. Metal grinding sounds. It is currently being emulated till replacement arrives tomorrow.
Regardless, same issue with connection. I have swapped ram, motherboard, and cpu. I have tried usb ethernet adapter and also installed a wifi card and still see the same results.
I setup a spare router with zero devices connected except unraid server and it still drops off randomly. Syslog isn't showing any network drops or link down messages.

With disk6 being emulated, how do I browse the files on that disk? I physically removed it from the machine.

If the emulated disk is mounting, you can still access the data under /mnt/disk#, if it's not mounting, please post the diags.

3 hours ago, JorgeB said:

If the emulated disk is mounting, you can still access the data under /mnt/disk#, if it's not mounting, please post the diags.

its not mounting

diags.zip

Jul 9 09:41:18 Tower kernel: XFS (dm-5): Corruption warning: Metadata has LSN (1:340126) ahead of current LSN (1:340120). Please unmount and run xfs_repair (>= v4.3) to resolve.

Jul 9 09:41:18 Tower kernel: XFS (dm-5): log mount/recovery failed: error -22

Check filesystem

5 hours ago, JorgeB said:

Jul 9 09:41:18 Tower kernel: XFS (dm-5): Corruption warning: Metadata has LSN (1:340126) ahead of current LSN (1:340120). Please unmount and run xfs_repair (>= v4.3) to resolve.

Jul 9 09:41:18 Tower kernel: XFS (dm-5): log mount/recovery failed: error -22

Check filesystem

After running with -L it’s mounting now

So what I want to do is going back to mostly fresh install of Unraid. I moved all files using unbalance to disks 1 & 2 and only want to keep these in the array for testing.

Since they all have no files, can I just do new config, remove them, and check parity is valid?

I need to keep previous apps, cache and array assignments. I don’t really need any other settings to be saved. What do I need to pull from the usb drive to do this?

7 minutes ago, jmztaylor said:

Since they all have no files, can I just do new config, remove them, and check parity is valid?

You cannot tick parity is valid as removing a drive automatically invalidates parity.

I removed all the HBA drives from the array.

I recreated unraid flash drive and only copied over drive assignments and key.

I am no longer having issues with the network. So I am going to assume something got messed up with the flash drive.

It's a strange one, but it's possible.

So everything seems to be going fine now that I rebuilt the flash drive, except mover. It is all over the place with speeds but mostly hang around 20-30MB/s which is really low. Not sure what could be causing it and I verified nothing is accessing the drives except mover as the drives were spun down before starting mover.

tower-diagnostics-20250718-1404.zip

Maybe you had Turbo Write enabled before?