I’ve had my switch and APs for about a week now… one thing I’ve noticed, that any host with more than one IP is causing the manage dashboard to sometimes report an IP, sometimes not. This causes issues when I go look for an IP in the portal, and it’s not there for the time it takes for whatever is capturing that info to change (the arp tables?)
1 Like
Thanks for your report. I’d like to investigate this further and gather some more intel that we can use to resolve this issue.
- Would you mind confirming if this happens with any specific devices exclusively?
- Does this occur on both wired and wireless devices?
- Could you check the same device in each scenario, such as a laptop with an ethernet port?
Hi!
#1). It’s my two servers running docker. So I’m using ipvlan, but the docker interfaces are on the same network segment. If I watch enough, my main server will swap IP addresses in the dashboard.
#2) Wired only. But at this point wireless is mostly just retail devices.
- not sure the ask on this one
I want to pop in and let you know I’m still looking into this and trying to confirm that we can reproduce what you’re seeing. I will fill you in when I have some confidence one way or another about that.
In the meantime, this is mostly just a cosmetic bug for you correct? I see how it may be frustrating for the data to be inconsistently displayed, but it doesn’t sound like it’s actively harming your network. Do I have that right?
Same for me @Alta-Josh and no, it’s not harming the network of course but rather a really annoying bug.
I tend to sort my clients by IP address and you can imagine what happens then…
Thanks for looking into this and even more for fixing this in an upcoming update 
1 Like
If there is any data I can provide, I’m happy to do so. I also sent some info via DM. It is not harming my network, but it pretty much negates being able to look for an IP in my network. I spent the first few days. Of my migration to Alta just trying to figure out what port this device was in. The manage dashboard needs to understand that the MAC to IP address mapping can be a one to many relationship. I don’t know how you are reporting that info manage, but after looking on the switch, perhaps you are feeding the info into the mqtt broker and it just grabs the latest Mac/IP pairing, thus showing only that moment in time?
1 Like
I appreciate the extra details. It gives me another element to investigate that I haven’t always thought of yet.
I’ll keep looking into it.
For me it’s wired devices only, and they all have one thing in common:
they get their IP address via DHCP reservation on my firewall.
So, fix IP but not defined on the device itself (still DHCP) but via MAC address reservation on the firewall. No matter which port on my S8-POE, no matter if POE provided or not, no other observations.
IP lease time either 14400 or 86400 secs.
Tested default VLAN1 only and with additional VLANs, no difference.
Between 5-10 devices affected, different vendors.
Does that help a bit?
1 Like
So I can repeat this very easily… From my Linux machine
ifconfig ens8:1 192.168.8.5
So I set an alias
Now I open up the manage dashboard and look for the primary address on the machine and find it
Then I ssh into another Linux host and
arping 192.168…8.5
I go and check primary IP in manage and it’s gone.
arping 192.168.8.200 (primary IP)
Look in manage and alias is now gone and it’s reporting the primary IP address again
I can now switch the IP reported for that port in manage on demand
1 Like
I have been able to see my Control report the secondary address, but I’m unable to reproduce the case where the IP vanishes entirely.
ip address add 192.168.1.244/24 dev enp0s31f6 label test:addr
Is there a way for you to look at the telemetry from my instance? Let me go and look at how ipvlan assigns the IP in Linux in the meantime…
How are you identifying the IP? I’m also seeing some people complain about this with ubiquity as well…
If the device on that port only uses the IP very infrequently, like less than once every five minutes, then its address will no longer be reported on the device list. We don’t support tracking multiple IPs per MAC currently, and will just report the most recently used IP, if it’s been used in the last five minutes.
If two IP addresses are required on the same interface, mapping it to an additional MAC address is recommended.
So that’s what causes the issue. Ipvlan in docker uses the same MAC address. Thought for future feature, would there be a way to pin a primary IP address in the metadata of a port… But that might not even help.
Please add a feature request in the community here. Sorry we couldn’t discover the root cause sooner, but thanks so much for your reproduction steps.