1 nic, 3 IP addresses and Unraid with NetAlertX running on Home Assistant.

[mattlward]

I have an Unraid server running PiHole and Asus Asrock IPMI and the main Unraid web gui running on 1 10g interface. Each has it's own IP address on the same sub-net, NetAlertX sees the PiHole address and the IPMI interface ip address but does not see the main server ip address.

All are pingable, all are in local DNS and all are up all the time.

So, my questions...

1. How do I configure the child/nic/virtual for the main server and the sub-servers? If it would find the main server I would define the sub-servers as virtuals under it, I would assume. But again, it does not find the main server at 192.168.1.100, it only finds the IPMI at 192.168.1.216 and the PiHole at 192.168.1.24. I have tried adding the main server and firing a single ICMP cycle at the subnet and it did not "find" 192.168.1.100 even though I can ping that address from all hosts on the subnet. Do I need to delete the sub-servers, add the main and then add them in as virtual servers?

Also, does this support V6?

[jokob-sk]

Hi @mattlward ,

You can always add a device manually and use the ICMP plugin to report its' status.

Unsure why the main server is not detected, but it usually boils down to either inaccessibility of the server from the docker container by the enabled scans, or a missing or incorrect interface configuration in the SCAN_SUBNETS setting. I would recommend the following:

1. Read the scan subnets docs: https://jokob-sk.github.io/NetAlertX/SUBNETS/
2. Read the scanning inaccessible networks guide: https://jokob-sk.github.io/NetAlertX/REMOTE_NETWORKS/
3. ssh into the container and execute the scans manually to see if the server is reachable/pingable
4. Last resort: add the server manually and use ICMP to report on its' status

I don't have V6 enabled, but some users seem to be using it without issues, although there might be some double reporting/duplication which I might need still to solve for: #1273

[jokob-sk]

192.168.1.100 d0:50:99:fb:bf:4c ASRock Incorporation
192.168.1.24 d0:50:99:fb:bf:4c ASRock Incorporation
Incorrect... the .216 shows up and the .24 did until I deleted it to try to add the .100 manually, now the .24 is not showing up and the .100 has never shown up. I really need/want the .100 to show up.

I see, thanks for the clarification. I think what is happening is that the application detects on the IPs and assigns it to the MAC address. If a subsequent IP is detected for an existing MAC then that will be ignored as MACs are used as unique IDs for Devices. Currently the MAC is a unique identifier, so to track different IPs on teh same MAC you will have to add them manually and track their status with the ICMP plugin.

[mattlward]

So, I have tired to add the same mac with the alternate IP and I am running the ICMP module as well. The device is never picked up, it feels like this is a problem. Fing Desktop also has the same issue. So, the .100 is the base interface for the server and the .24 is a service bridged on to the same ethernet. I will try adding it again, are there any checkboxex that I need to check to make this work? In a perfect world I would have the .100 listed as a direct node off of the switch and then the .24 would be shown as a child off of the .100.

I still have not been able to figure out how to get output to HA in the form of device status and numbers of up/down or sideways machines. I have MQTT enabled and I think Mosquito is connected.

I love it, I am a senior network engineer by trade for 35 years, but I am not a software network engineer so I know one side really well but not the other. Great package... Any chance of using LLDP in the future to learn some of the network mapping? I do you switches that support LLDP and have agents on all systems that support it and things like my APs support it as well.

[jokob-sk]

So, the .100 is the base interface for the server and the .24 is a service bridged on to the same ethernet.

I'm not that great with network configurations (I know, the irony - partly why I got into this project to learn a bit more) so not sure if I follow, but if other tools have a similar problem than it might be an network environment thing?

The ICMP plugin only pings IPs that are availble in NAX, so if a device can't be pinged then not sure how to report on it (maybe some other plugins might help?) So just to verify, I'd:

1. Create a device for 192.168.1.100 manually (e.g MAC d0:50:99:fb:bf:4c)
2. Create a device for 192.168.1.24 manually (e.g MAC 20:50:99:fb:bf:4c) (different from above)
3. Enable ICMP and limit the IPs to scan (ICMP_IN_REGEX) to 192.168.1.100|192.168.1.24

For MQTT troubleshooting, check these:

1. https://jokob-sk.github.io/NetAlertX/HOME_ASSISTANT/?h=home
2. https://github.com/jokob-sk/NetAlertX/tree/main/front/plugins/_publisher_mqtt#forcing-an-update
3. How to know if a device is down (vs offline) in Home Assistant? #1057

Any chance of using LLDP in the future to learn some of the network mapping?

Please open a separate feature request and I will have a closer look in the future - not sure about the complexity so can't say.

I love it, I am a senior network engineer by trade for 35 years, but I am not a software network engineer so I know one side really well but not the other. Great package...

Glad to hear that! :) I love when different people with different passions and expertise overlap - give the project a shout out on socials if you can - I dream of this becoming my source of income (Somehow, not sure how :D ) in future, so the more people use NAX, the more likely I can focus on this more instead my 9-5 :)

[mattlward]

So, the .100 is the base interface for the server and the .24 is a service bridged on to the same ethernet.

I'm not that great with network configurations (I know, the irony - partly why I got into this project to learn a bit more) so not sure if I follow, but if other tools have a similar problem than it might be an network environment thing?

So, no the network is ok... the issue is that some tools do not understand bridging and common mac addresses for different IP addresses. This can cause issues if one runs spanning tree on a network port serving the host. I am not sure if I misspoke or was not clear earlier... both the .100 and .24 are pingable from the cli within the container.

The ICMP plugin only pings IPs that are availble in NAX, so if a device can't be pinged then not sure how to report on it (maybe some other plugins might help?) So just to verify, I'd:

1. Create a device for `192.168.1.100` manually (e.g MAC `d0:50:99:fb:bf:4c`)

2. Create a device for `192.168.1.24` manually (e.g MAC `20:50:99:fb:bf:4c`) (different from above)

3. Enable ICMP and limit the IPs to scan (`ICMP_IN_REGEX`) to `192.168.1.100|192.168.1.24`

So, I got it finally. I had to assign a bs mac address to the main server and let ICMP find it and then defined it as a switch, then i added the .24 and made it a nic on the .100. Not sure if it should be a virtual or a nic.

I do like it, I do however wish it could handle this properly or gracefully. Still not sure I have the relationship set up correctly, but the only way I could find to tie the 2 together was to make the .100 (main server) a device that I could attach other things to.

Matt

[jokob-sk]

Great - glad to hear it's working :)