You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I just noticed that after a change reorganisation, where I am testing out an idea and Technetium DNS from a whole VM to itself (with Docker) on a dedicated VM server. To running in Docker, within an LXC container, under Proxmox (I know that sounds horrid, but it's actually faster and uses less hardware and very disk/memory efficient). It now runs in just 512k of RAM and 2G HDD and boots from cold in 2-3 seconds, With the option of moving it's actual execution location to one of four other machines, in the event of something breaking.
I set this running yesterday but over time started seeing machines struggling to connect to the internet. But of my own three machines only one had this issue.
Anyway I just noticed it was offering it's container (non routable) address as the address for DNS services to clients in their DHCP leases, rather than the address it is told to offer to clients in the setting "Use this address to identify itself".
As a workround, I updated each DHCP scope and unchecked the 'use own address' option and manually enter the correct address. Did not take long to fix but did mean very odd behaviour for a few hours as there is a backup DNS that also knows about this one, so it was still getting queries, just the long way round and was getting a lot of random DNS timeouts which I thought was the ISP as it 'worked for me'.
Note sure what the intended behaviour is but as the listing page for DHCP scopes does not include the DNS, it had me confused for a while as if I made test queries (unintentionally) using the correct address there seemed to be no problem. Maybe make more fields viewable or change the way it determines it's own address?
The text was updated successfully, but these errors were encountered:
Thanks for the post. The option to automatically select this DNS server for configuring DHCP option will use the IP address that is available on the system. Since you are using the DNS server in a container, it will pick up the IP available on the container which may not be routable. So, you will need to manually configure the desired IP address in such scenario.
Anyway I just noticed it was offering it's container (non routable) address as the address for DNS services to clients in their DHCP leases, rather than the address it is told to offer to clients in the setting "Use this address to identify itself".
Regarding the option "Use this address to identify itself", I do not see any such option anywhere. Which specific option are you taking about here?
Note sure what the intended behaviour is but as the listing page for DHCP scopes does not include the DNS, it had me confused for a while as if I made test queries (unintentionally) using the correct address there seemed to be no problem. Maybe make more fields viewable or change the way it determines it's own address?
The DHCP scope edit page does list the DNS Server option clearly. Are you talking about some other option?
I just noticed that after a change reorganisation, where I am testing out an idea and Technetium DNS from a whole VM to itself (with Docker) on a dedicated VM server. To running in Docker, within an LXC container, under Proxmox (I know that sounds horrid, but it's actually faster and uses less hardware and very disk/memory efficient). It now runs in just 512k of RAM and 2G HDD and boots from cold in 2-3 seconds, With the option of moving it's actual execution location to one of four other machines, in the event of something breaking.
I set this running yesterday but over time started seeing machines struggling to connect to the internet. But of my own three machines only one had this issue.
Anyway I just noticed it was offering it's container (non routable) address as the address for DNS services to clients in their DHCP leases, rather than the address it is told to offer to clients in the setting "Use this address to identify itself".
As a workround, I updated each DHCP scope and unchecked the 'use own address' option and manually enter the correct address. Did not take long to fix but did mean very odd behaviour for a few hours as there is a backup DNS that also knows about this one, so it was still getting queries, just the long way round and was getting a lot of random DNS timeouts which I thought was the ISP as it 'worked for me'.
Note sure what the intended behaviour is but as the listing page for DHCP scopes does not include the DNS, it had me confused for a while as if I made test queries (unintentionally) using the correct address there seemed to be no problem. Maybe make more fields viewable or change the way it determines it's own address?
The text was updated successfully, but these errors were encountered: