I’ve found an odd bug that I’ve been able to replicate over the two devices that I have now (7249 and 7289). I originally had these devices setup using the default DNS entry because I was originally using the TTN network. However, I recently setup my own chirpstack that is local to our environment, and our DNS entry for that device is not open to the public. I then set the DNS servers on these devices to use our local campus DNS server so it could resolve the hostname of my chirpstack server. However, I found that despite me manually configuring these DNS servers via the GUI, the RAK devices weren’t using those DNS servers and was unable to resolve the address of my chirpstack server. To get around for this for now I’ve ssh’d into the devices and manually set the /etc/resolv.conf file with our DNS server information which fixed my issue. However, this file does get reverted back if the device reboots so it’s not sustainable. Has anyone switched dns servers in use and gotten the GUI to actually honor the set dns servers?
It’s not our Gateway but a customer is reporting the issue, so I can’t provide too many details. But it sounds exactly as what OP has stated. The information I got from the customer is:
They want to migrate to some internal infrastructure. For that, they have set some private DNS Servers (10.x IP Address)
Resolving the DNS entry of the MQTT Endpoint is not working with that setup. However, if they put the DNS Server manually in the /etc/resolv config file on the Gateway (They accessed it via SSH) DNS resolve of the MQTT Endpoint works fine
However /etc/resolv is overwritten at bootup this is not a permanent solution
I asked them to contact the RAK Wireless Support themselves, maybe you’ll get a support query with more details as I wasn’t able to reproduce the same with our Gateway (Didn’t have a private DNS server to test though)
No, not any real fix. We did setup a script to constantly check and update the /etc/resolv.conf. I had sent several inquiries to RAK about the issue via their website, but they never responded to them. So far every RAK gateway I’ve purchased (about 6 of them now has had this issue).
For the RAK rep in this thread here’s the issue we had:
Our network/application server (chirpstack) is not resolvable off-campus and only via our local DNS servers. However, if I use the GUI to set the DNS servers on the RAK gateways to our local DNS servers it does nothing. Looking at the logs, I can see resolution of our chirpstack server is not happening. The only way to resolve this, is by ssh’n into the gateway and manually changing the /etc/resolv.conf. As stated earlier though, this is only a temp bandaid workaround as this file gets wiped and replaced with 127.0.0.1 after a reboot / power cycle.
Hi Nikola, for now the customer solved this by placing a fixed entry in the /etc/hosts file. However not a really good permanent solution. They have contacted the [email protected] but have not received any feedback yet, can you maybe check that? They can for sure share the requested details with you. Thanks a lot
We can not find the request. Can you tell me email of the customer (on direct message) or the company details or something so I can find the request and contact the customer directly?
I found one communication with you on the topic of the DNS, where our staff member replied to you, but you never answered the email.
The V1 hardware is end of life, but if you have a V2 gateway with the problem, please contact us at [email protected] . Also, check your spam folder; maybe you lost our response there