[BUG]: no access 172.28.0.0/16 after system_server restarted

[libdmmy]

Kernel Version

4.19.157 (recompiled with required options)

Kernel Source Link (REQUIRED)

https://github.com/Havoc-OS-OnePlus-8-Series/kernel_oneplus_sm8250

Droidspaces Version

6.4.5

Rooting Method

Magisk

Device OEM & Model

OnePlus 9R

Android Version & ROM

HavocOS (A11)

Execution Mode

DAEMON

Networking Mode

NAT

Describe the Bug

after system_server restart (soft reboot) network requests to containers from host seem to be not handled until I restart containers

Steps to Reproduce

1. launch container with 172.28.1.1
2. ssh into 172.28.1.1 works fine
3. su -c killall system_server
4. connecting with ssh does nothing

Logs / Screenshots

No response

Required Acknowledgments

• I have verified this is not a duplicate issue.
• I am using a kernel compiled strictly according to Droidspaces' official documentation.
• I am NOT using a kernel with 69+ random configs, CRC nukes, or a broken ABI.
• I have confirmed this issue persists in both Daemon and Direct modes.
• I admit that I am a clown for checking this box, confirming I have NOT read these rules.
• I admit I did zero research, didn't ask my AI waifu for a fix, and am dropping this here because I'm 100% sure the fault lies with Droidspaces.

[ravindu644]

Thank you for your bug report. Confirmed by reproducing the exact behavior you've described. I'll work on fixing this.

a16xm:/ # droidspaces show                                                                                                            

┌──────────────┬──────────┐
│ NAME         │ PID      │
├──────────────┼──────────┤
│ Ubuntu-24.04 │ 23679    │
└──────────────┴──────────┘

a16xm:/ # droidspaces info                                                                                                            
Host: Android aarch64

Container: Ubuntu-24.04 (RUNNING)
  PID: 23679
  OS: Ubuntu 24.04.4 LTS
  Hostname: ubuntu
  Uptime: 1d 15h 38m 20s

Features:
  Networking: NAT
  NAT IP: 172.28.234.160
  SELinux: Enforcing

a16xm:/ # droidspaces run ping -c 2 google.lk                                                                                         
PING google.lk (172.217.194.94) 56(84) bytes of data.
64 bytes from si-in-f94.1e100.net (172.217.194.94): icmp_seq=1 ttl=106 time=92.4 ms
64 bytes from si-in-f94.1e100.net (172.217.194.94): icmp_seq=2 ttl=106 time=96.3 ms

--- google.lk ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1002ms
rtt min/avg/max/mdev = 92.406/94.371/96.337/1.965 ms
a16xm:/ # ip route show default                                                                                                       
172.28.0.0/16 dev ds-br0 proto kernel scope link src 172.28.0.1 
192.100.100.0/24 dev wlan0 proto kernel scope link src 192.100.100.100 

# triggering soft reboot
1|a16xm:/ # killall system_server

# during soft reboot
a16xm:/ # ip route show default                                                                                                       
172.28.0.0/16 dev ds-br0 proto kernel scope link src 172.28.0.1 
192.100.100.0/24 dev wlan0 proto kernel scope link src 192.100.100.100 linkdown 
a16xm:/ # ping -c 2 google.lk
ping: unknown host google.lk

# adb access lost during soft reboot
2|a16xm:/ # ravindu644@fedora:~$ adb shell
adb: no devices/emulators found

# after soft reboot, android started again
ravindu644@fedora:~$ adb shell
a16xm:/ $ su
a16xm:/ # ip route show default
172.28.0.0/16 dev ds-br0 proto kernel scope link src 172.28.0.1 
192.100.100.0/24 dev wlan0 proto kernel scope link src 192.100.100.100 
a16xm:/ # ping -c 2 google.lk
PING google.lk (142.250.197.35) 56(84) bytes of data.
64 bytes from nchkga-ag-in-f3.1e100.net (142.250.197.35): icmp_seq=1 ttl=113 time=105 ms
64 bytes from nchkga-ag-in-f3.1e100.net (142.250.197.35): icmp_seq=2 ttl=113 time=124 ms

--- google.lk ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1002ms
rtt min/avg/max/mdev = 105.729/114.989/124.250/9.266 ms

# container is still running
a16xm:/ # droidspaces show                                                                                                            

┌──────────────┬──────────┐
│ NAME         │ PID      │
├──────────────┼──────────┤
│ Ubuntu-24.04 │ 23679    │
└──────────────┴──────────┘

# even though we have a healthy default route, and
# working ping, droidspaces internet is dead
a16xm:/ # droidspaces enter                                                                                                           
root@ubuntu:~# ping google.com -c 4

^C
root@ubuntu:~# 

[libdmmy]

uh, I mean I can't connect to a container (like ssh into it) at all

[ravindu644]

uh, I mean I can't connect to a container (like ssh into it) at all

the entire container NAT is dead, resulting dead in SSH

[ravindu644]

Fixed your issue in here:

cf9f008

please check the https://t.me/DroidspacesCI channel once the new CI apk has been built

[ravindu644]

here it is:

https://t.me/DroidspacesCI/671

https://github.com/ravindu644/Droidspaces-OSS/actions/runs/31293900849/artifacts/9032372060

[libdmmy]

can confirm it's fixed now