For some reason my TP-Link router ER-605 V2.0 out of nowhere started complaining about maxed out log.
I checked and cleared them and thought nothing of it, as it was the first time since I sat the system up, a year ago or so.
Just a week or 2 after I got the same message again when I logged into my Omada setup, log was full etc.
The log was full of devices getting assigned IP adress 192.168.1.xx
DHCP Server allocated IP address 192.168.1.xx for the client[MAC:xxxxxx ]
I then realised all of the events logged was from LXC containers on my Proxmox constantly refreshing their IP. Not getting new, just dropping out, then getting the same IP given again.
So I have 9 containers, 8 of them are refreshing their IP. All of the containers are set up by the VE helper Scripts. This refreshing of the IP’s happening like every 15-20 min.
IP lease time is set to 24 hours.
I also have 2 VM’s running on the Proxmox as well, no problem there. (1x win10 & 1x win11).
So with 2 VM’s and 1 LXC container not doing this, I kinda rule out physical connectivity issues.
I tried setting a static IP from within Omada to one device, but that did not stop it from dropping IP and getting it back again.
I assume you have the LXC containers configured to pull an IP address from DHCP instead of assigning them a static one, given the problem behavior. However, is that your intended configuration? You could assign them a static one under the container > network > click the interface > Edit.
I tried setting static IP, both from Omada setup and from Proxmox. (Not at the same time). No change in behavior.
But I noticed a repeating pattern.
After some time the LXC drops of the network. Then when I open the page in my browser to access said service. It reappears in my list of devices on the overview of the Omada setup.
It seems like the LXC is going in to some sort of “sleep mode” and drops out of the network. And then" wakes up" when it gets a “request” or something.
This can explain why 1 LXC does not show this behavior, as it is a LXC for changedetection.io, and is set up to check certain sites for updates frequently, and therefore never goes offline.
Other containers are Stirling PDF, n8n, trilium, Homebox, Homepage etc.
Seems like after a set amount of time the containers send a “ping” of some sort to tell the network that they are still there. And every time they do the router assigns the IP adress to them again and creates an event in the log. So either they need to be “online” to stop dropping out of network, or they need to send this “ping” at shorter intervals…
I have turned off the logging of said event right now, to stop the Omada setup getting flooded with event logs . But is this normal behaviour from the LXC installs? Is there a way to change this?