Networking Problems - Host unable to access internet, VMs and bridged dockers fine

[hellarda]

Hi all,

 

Have been away for a few weeks so not been on my server for a while. Just got back and everything on the host interface inc unraid itself can't seem to access the internet. 

 

Nothing has changed network wise, and DNS is still set correct to 8.8.8.8.

 

Dockers and VMs using br interfaces can access the internet fine. Can ping my router from the the cmd line but can't ping 8.8.8.8

 

Diags attached. Any clues?

 

Thanks

hellard-server-diagnostics-20190715-2209.zip

Edited July 17, 2019 by hellarda

[Frank1940]

This may be a very stupid question but is your router aware of the static IP addresses of your Unraid server?  (After several weeks of non-use, any DHCP leases to missing devices (turned-off or on the trip) would have expired.  These previous clients might have had new addresses assigned to them after your return creating a conflict. )    

[hellarda]

Thanks for the reply - no my DHCP pools sit way outside the static IPs I use

[Frank1940]

I run a very simple Network setup on my server but you seem to have (what seems to me) a very complicated one.  Have you looked at your Diagnostics file?  Specifically, /config/network.cfg     /config/network-rules.cfg   /system/ifconfig.txt     and    /system/ethtool.txt   to see if this is how you want to have the networking setup.  You also only have DNS service setup for eth0 which is for "Management".   

 

You might also want to change/edit your title for this thread to indicate better what your networking problem is.  If it possible indicate which subsystems are having problems-- Unraid itself (including plugins), Dockers or VM's.  You can do this by editing your first post.  

Edited July 17, 2019 by Frank1940
Forgot a file

[hellarda]

There is only the option to add DNS to the first interface

 

From the config files it does seem to be complicated. However looking at the config in the GUI it looks simple.

 

My network setup should be:

 

Management Interface eth0 10.23.0.11

2nd Management Intrface eth1 10.23.0.12

Docker br interface 10.23.0.13

VM br interface 10.23.0.14

 

Which is what I have setup in the GUI.

 

For some reason ping requests from the terminal respond from 10.23.0.17, which in the logs is assigned to docker0

 

I really don't understand whats going on!! 

 

I can't access community Apps, and can't search for OS updates from the GUI

 

I don't understand how this has just suddenly changed without any config changes from myself....

[Frank1940]

Neither do I!  As I said, I would be editing that first post.  Can you reach the Internet from the VM?  Can you reach the Internet from Dockers?  Does File sharing work?  Can you access the GUI?  Once you have define the total problem, you should be able to modify that title to get the real Gurus in that area looking at your problem!  The present title is just too broad.  (I am more of an SMB expert than anything else...)

[Arubial]

what's the default gateway set to for the host NIC? (it should be your router IP)

[ken-ji]

your biggest mistake is assigning 8 IPs to Unraid on the same physical network.

This will make networking work in ways you will not predict or understand.

What you probably want here is to have just two bridges, and only br1 has and IP (10.23.0.11/24 - gateway 10.23.0.1)

put eth0, eth1, eth2, eth3 together and bonded and bridged to br0, and assign the desired IP here

then put eth4, eth5, eth6, eth7 together as bonded and bridge br4 (i think this is the correct one, else it would be br1)

Configure the docker network pool to custom and delete the default one to br0, and create one for 10.23.0.0/24 (or smaller) on br4

point your containers to this network interface

link your VMs to either bridge. 

that will simplify your life and make your network easy enough to understand:

unraid is reachable via the first bond/bridge

dockers on 2nd bond/bridge

VMs on either it is connected.

[hellarda]

On 7/18/2019 at 1:55 AM, ken-ji said:

your biggest mistake is assigning 8 IPs to Unraid on the same physical network.

This will make networking work in ways you will not predict or understand.

What you probably want here is to have just two bridges, and only br1 has and IP (10.23.0.11/24 - gateway 10.23.0.1)

put eth0, eth1, eth2, eth3 together and bonded and bridged to br0, and assign the desired IP here

then put eth4, eth5, eth6, eth7 together as bonded and bridge br4 (i think this is the correct one, else it would be br1)

Configure the docker network pool to custom and delete the default one to br0, and create one for 10.23.0.0/24 (or smaller) on br4

point your containers to this network interface

link your VMs to either bridge. 

that will simplify your life and make your network easy enough to understand:

unraid is reachable via the first bond/bridge

dockers on 2nd bond/bridge

VMs on either it is connected.

Thanks for the reply - I've always found unraid networking a little tricky to understand. I did actually disable 4 of my NICs a while ago to try and simply the whole thing (and they don't show up in the GUI settings page anymore, but looks like they are still doing something?) 

 

I'm struggling to follow your advise above, are you able to point in the right direction of achieving this?

 

Thanks

[hellarda]

Right I've got a bit further I think....!

 

It seems the host was trying to route through br7 randomly (an old bridge interface that I had removed all interfaces from, but apparently hadn't deleted the bridge itself?)

 

So having done ifconfig br7 down, it's gotten rid of that interface.

 

Now when I ping from terminal I get a reply saying network unreachable, rather than a host unreachable reply from the br7 address.

 

Presumably I need to add a route back in, in order to get everything to route through the correct interface again, which I think I've done, but still having no joy.

 

Attached a screenshot of my routing table. Also attached is screenshots of the br configs.

 

Any ideas?

 

Thanks

[hellarda]

More developments... 

 

Managed to delete br7 and br6 and setup a default route for br0. Great.

 

Reboot - br7 and br6 come back, default route disappears ☹️

 

I'm doing all these changes in terminal. Any ideas how to make them stick?

[hellarda]

Is anyone able to help shed light on this? br7 and br6 come back after a reboot and I have no idea why?

[ken-ji]

Reboot and grab the diagnostics