DNS in Ubuntu 22.04.5 LTS

[doanvtamhuynh]

I ran hev-socks with the following configuration:

tunnel:
  # Interface name
  name: tun0
  # Interface MTU
  mtu: 8500
  # Multi-queue
  multi-queue: false
  # IPv4 address
  ipv4: 198.18.0.1
  # IPv6 address
  ipv6: 'fc00::1'
  # Post up script
# post-up-script: up.sh
  # Pre down script
# pre-down-script: down.sh

socks5:
  # Socks5 server port
  port: *port-proxy
  # Socks5 server address (ipv4/ipv6)
  address: *address-proxy
#  address: 192.168.1.239
  # Socks5 UDP relay mode (tcp|udp)
  udp: 'udp'
  # Override the UDP address provided by the Socks5 server (ipv4/ipv6)
# udp-address: ''
  # Socks5 handshake using pipeline mode
# pipeline: false
  # Socks5 server username
  username: *'username-proxy'
  # Socks5 server password
  password: *'password-proxy'
  # Socket mark
  mark: 438

mapdns:
  # Mapped DNS address
  address: 198.18.0.2
  # Mapped DNS port
  port: 53
  # Mapped IP network base
  network: 240.0.0.0
  # Mapped IP network mask
  netmask: 240.0.0.0
  # Mapped DNS cache size
  cache-size: 10000

misc:
  # task stack size (bytes)
# task-stack-size: 86016
  # tcp buffer size (bytes)
# tcp-buffer-size: 65536
  # udp socket recv buffer (SO_RCVBUF) size (bytes)
# udp-recv-buffer-size: 524288
  # number of udp buffers in splice, 1500 bytes per buffer.
# udp-copy-buffer-nums: 10
  # maximum session count (0: unlimited)
# max-session-count: 0
  # connect timeout (ms)
  connect-timeout: 30000
  # TCP read-write timeout (ms)
  tcp-read-write-timeout: 300000
  # UDP read-write timeout (ms)
  udp-read-write-timeout: 60000
  # stdout, stderr or file-path
  log-file: stdout
  # debug, info, warn or error
  log-level: debug
  # If present, run as a daemon with this pid file
# pid-file: /run/hev-socks5-tunnel.pid
  # If present, set rlimit nofile; else use default value
# limit-nofile: 65535

and configure the IP rule as follows:

sudo ip rule add fwmark 438 lookup main pref 10
sudo ip route add default dev tun0 table 20
sudo ip rule add lookup 20 pref 20
sudo ip -6 rule add fwmark 438 lookup main pref 10
sudo ip -6 route add default dev tun0 table 20
sudo ip -6 rule add lookup 20 pref 20

After performing all the steps above, the IP address has been changed to the proxy server's IP address, but the DNS remains at the Hong Kong DNS server address instead of the DNS USA server, as shown in the image below:

Do I need to add any configuration or make any adjustments to get the DNS to work correctly with the server? Or is it a bug when using hev-socks on Ubuntu?

[heiher]

Set the mapdns.address same as the dns server in /etc/resolv.conf.

[doanvtamhuynh]

It worked correctly, thank you for the instructions.

[doanvtamhuynh]

Can I ask why I'm getting authentication errors when using that configuration on arm64?
It works when using a proxy without authentication, but when I have authentication, I lose connection.

[heiher]

Which auth type are you using?

[doanvtamhuynh]

username,password:

socks5:
  port: *my-port-proxy
  address: *my-address-proxy
  udp: 'udp'
  username: *'my-username-proxy'
  password: *'my-password-proxy'
  mark: 438

[heiher]

Could you capture network packets contains auth with curl(succ) and this client(fail)?

[doanvtamhuynh]

I used the configuration as above. With a proxy that doesn't require a username and password, everything works fine, but when using a proxy that requires username and password authentication, it doesn't.

[heiher]

So I need the packets to see what happened

[doanvtamhuynh]

this is tcpdump log:
log_tcpdump_hevsocks.txt