Hello,
I operate a dense Unifi network at a school with 56 access points (AP-AC Pro and nano HD) behind an UDM Pro and 17 Unifi switches. Each room is equipped with an AP, so that in principle an AP does not have to supply more than 5-7 meters in a radius “alone”.
About half a year ago the WIFI was consolidated - before we only had about half of the access points. Only every 2nd room was supplied with an AP.
Since the upgrade, I’ve noticed that wireless devices have a good WiFi signal, but no website is reachable (DHCP?). But push notifications do arrive …??
A short deactivation and reactivation of the WIFI on the device solves the problem.
I have already tried a lot of settings on the controller and also tested various firmware updates. Unfortunately, this problem remains unsolved. So far, we have not changed anything in the wireless settings of any individual AP. The standard values are set here.
However, after the densification, we initialized the RF scans on 2G and 5G and initiated the automatic adjustment of the channels.
Maybe you have a suggestion?!
Thanks and greetings from Germany (and sorry for bad english)
Daniel
HI there, could this be a sticky mac issue? Are your device trying to hold on as long as they can to the AP they originally associate with? What is the received signal strength from a neighboring room AP, and what is your lowest data rate you have enabled?
On the other side of the coin, I know it might be a pain, but what happens if you temporarily disable every second AP in the attempt to restore the RF environment back to pre upgrade - see if your connectivity restores?
Hi. thanks for your replies. @Pete I don’t think, this is a Mac-Issue. It happens at random intervals on many different clients. I also thought that the device might try to keep the connection to the original AP as long as possible. That’s why I set the radio transmit power on all APs to “low”. The problem still persists. The client even shows full signal when the problem occurs. I suspect a timeout, roaming or power-safe issue. The client also has a correct IP address (@LTS_Tom ).
In anomalies the message “Client is having trouble resolving a domain name to an IP address (DNS timeout)” is often to read.
Hi. I actually manage a similar Ubiquiti - UniFi setup in a school.
Issues I’ve run into that can cause this:
loop in the network (check for duplicate IPs in the MAC table on a Ubiquiti switch)
bad DNS config(if the devices are hitting a bad DNS server it can cause severe delays and even break the connection)
I’d suggest checking UniFi Controller for if your APs possibly have a different DNS configuration. Cannot comment on how to do all this via UNMS as I have yet to try it.
Thanks @vaughng fpr the reply - the APs DNS ist google 8.8.8.8, the WAN DNS is a Pihole with a local IP - ist it a problem, whenn the DNS of AP and WAN are different?
Shouldn’t be an issue but I usually set them the same as the gateway DNS or to the gateway(our gateway is the firewall which will resolve internal queries and forward requests to the desired addresses). Your devices should be getting DNS servers from the gateway anyway though. You can check this on android devices by going to the WiFi network and looking at the details.
To check MAC tables on the new EdgeSwitch interface, log in to the switch and on the left menu bar click the globe icon(the round www internet icon). Look for duplicates, if there is a loop there will be lots of duplicates when there are users connected to the wireless.
I would think it has something to do with piHole though. Where are you running it? Does it have enough resources?
Yes the Pihole has enough ressources. It runs on a VSphere Hypervisor. MAybe the Pihole is the Problem. But checking the Mac - tables are a good suggestion. Unfortunately i habe unifi switches - not edge switches… i think i have to check that in the controller
Unfortunately I can’t help you with the UniFi switches. However to test if piHole is the issue you can connect to the WiFi network and test the response times by pinging an external website or if you’re up to it look into DIG to test further. I’ve heard of many people running into issues with response time on piHole due to some minor configuration or the other.