joedotmac Posted September 10, 2018 Share Posted September 10, 2018 Experiencing inconsistent results from configuring UNRAID with bond interface and a standalone management interface both on same LAN 192.168.199.0/24 Bond0 (br0) MAC Addr 29:30 IP 192.168.199.111 (eth0, eth1, eth2) eth4 MAC Addr 29:34 IP 192.168.199.100 Seems the configuration of network interfaces with associated interfaces as indicated in the GUI are not consistent with whats being advertised on the layer 2 data link layer. From a client machine on the same 192.168.199.0/24 LAN, displaying the arp cache will indicate both the .111 and .100 IP's are using the same MAC address of 29:30. Then I can initiate a file transfer to .111 and it will use eth3 and not the expected bond interface. Are there any files an I inspect or methods of which force compliance of MAC address being advertised to the switch out of a physical ethernet interface? Quote Link to comment
JuliusZet Posted February 20, 2019 Share Posted February 20, 2019 I have a similar issue, probably related to the same bug in unRAID. Have a look at the following trace captured from my router: I have one management interface for unRAID itself (192.168.10.10) and 9 interfaces to be used by VMs (no IP addresses). However every time my router sends out an ARP request to find out the MAC address for the unRAID management interface, every interface responses. Only one of these responses has the right MAC address in it. So at the moment I can only reach my unRAID web GUI und network shares if I'm lucky ... I could not find out yet why this happens. Maybe someone else has any clues? Quote Link to comment
ken-ji Posted February 20, 2019 Share Posted February 20, 2019 On 9/11/2018 at 5:11 AM, joedotmac said: Experiencing inconsistent results from configuring UNRAID with bond interface and a standalone management interface both on same LAN 192.168.199.0/24 Bond0 (br0) MAC Addr 29:30 IP 192.168.199.111 (eth0, eth1, eth2) eth4 MAC Addr 29:34 IP 192.168.199.100 Seems the configuration of network interfaces with associated interfaces as indicated in the GUI are not consistent with whats being advertised on the layer 2 data link layer. From a client machine on the same 192.168.199.0/24 LAN, displaying the arp cache will indicate both the .111 and .100 IP's are using the same MAC address of 29:30. Then I can initiate a file transfer to .111 and it will use eth3 and not the expected bond interface. Are there any files an I inspect or methods of which force compliance of MAC address being advertised to the switch out of a physical ethernet interface? I could be wrong but AFAIK this is the default behavior for Linux machines with multiple IP addresses on different interfaces in the same subnet. 6 hours ago, JuliusZet said: I have a similar issue, probably related to the same bug in unRAID. Have a look at the following trace captured from my router: I have one management interface for unRAID itself (192.168.10.10) and 9 interfaces to be used by VMs (no IP addresses). However every time my router sends out an ARP request to find out the MAC address for the unRAID management interface, every interface responses. Only one of these responses has the right MAC address in it. So at the moment I can only reach my unRAID web GUI und network shares if I'm lucky ... I could not find out yet why this happens. Maybe someone else has any clues? How are your interfaces configured? I have only two ports with a similar configuration - only one management IP on br0 and no other IPs on br1 - and my managed switch has never seen the wrong mac address on the other interface. Quote Link to comment
JuliusZet Posted February 21, 2019 Share Posted February 21, 2019 13 hours ago, ken-ji said: How are your interfaces configured? I have only two ports with a similar configuration - only one management IP on br0 and no other IPs on br1 - and my managed switch has never seen the wrong mac address on the other interface. Thanks for your reply! I attached my diagnostics zip file, that includes information about my interface configuration. unraid-server-diagnostics-20190221-0853.zip I assume the issue is related to a bug in unRAID or in the driver for the network cards. Quote Link to comment
ken-ji Posted February 24, 2019 Share Posted February 24, 2019 I can only guess here, but are all the VM interfaces on the same physical network? Quote Link to comment
JuliusZet Posted February 25, 2019 Share Posted February 25, 2019 22 hours ago, ken-ji said: I can only guess here, but are all the VM interfaces on the same physical network? Yes, everything is on the same network. This is my network topology: About my future plans to work around this issue: I bought myself a new router that can handle multiple networks. (Believe it or not, my current router can not do this ...) I want every interface to have its own network. I will set everything up next weekend. Let's see if this resolves the issue. Thank you very much for your assistance by the way! Very much appreciated! :) Have a great day! Quote Link to comment
ken-ji Posted February 25, 2019 Share Posted February 25, 2019 (edited) that's why you are seeing the arps on all the interfaces. the easiest fix for you was to bond the interfaces on the same network together as depending on your router, may support link aggregation giving you the effect of 10g if you have many clients. Since you have a new router with VLAN support(?) You'll need to bond interfaces into groups that make sense to you then assign the appropriate VLAN (subnet) to the bonded interfaces. Most consumer routers do not support VLANs or multiple networks. The current user favorites seem to (in no order) pfsense, ubiquiti, and mikrotik(my choice as they are relatively cheap and easy enough once you understand it) Edited February 25, 2019 by ken-ji Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.