MacOS / Random network disconnect (Host-only Adapter) - ARP: duplicate use of $IP detected

[TB-Quentin]

Hi VirtualBox Use/Admin/Dev,

We are using VirtualBox in my company to launch Alpine Linux for Docker usage. All developers use it for development and the main application in Docker is Apache, MySQL and PHP. (The network is Host-only Adapter)

It's work really fine since several years but since 6 months peoples on MacOS suffer of random network disconnect. That lasts 2 to 3 minutes and it comes back without any intervention ...(This happens between 2 to 6 times over a period of one hour)

Jeu 14 nov 2019 13:22:40 EST: 64 bytes from 172.21.21.5: icmp_seq=6644 ttl=64 time=0.439 ms
Jeu 14 nov 2019 13:22:41 EST: 64 bytes from 172.21.21.5: icmp_seq=6645 ttl=64 time=0.296 ms
Jeu 14 nov 2019 13:22:42 EST: 64 bytes from 172.21.21.5: icmp_seq=6646 ttl=64 time=0.369 ms
Jeu 14 nov 2019 13:22:44 EST: Request timeout for icmp_seq 6647
[...]
Jeu 14 nov 2019 13:25:07 EST: Request timeout for icmp_seq 6789
Jeu 14 nov 2019 13:25:07 EST: 64 bytes from 172.21.21.5: icmp_seq=6790 ttl=64 time=0.832 ms
Jeu 14 nov 2019 13:25:08 EST: 64 bytes from 172.21.21.5: icmp_seq=6791 ttl=64 time=0.412 ms

First we check on the Guest host:

• Check network/application logs
• Update OS and packages
• Allocate more memory
• ....

After we make somes changes on Host:

• Update MacOS from 10.14 to 10.15
• Disable Bluetooth
• ...

Here the Host IP/vboxnet0:

MacBook-Pro:platform-website turbulent$ ifconfig vboxnet0
vboxnet0: flags=8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500
	ether 0a:00:27:00:00:00 
	inet 172.21.21.1 netmask 0xffffff00 broadcast 172.21.21.255

Here the Guest IP/eth1:

substance@work:~$ ifconfig eth1
eth1      Link encap:Ethernet  HWaddr 08:00:27:DA:BC:C4  
          inet addr:172.21.21.5  Bcast:0.0.0.0  Mask:255.255.255.0
          inet6 addr: fe80::a00:27ff:feda:bcc4/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:96227 errors:0 dropped:0 overruns:0 frame:0
          TX packets:68725 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:9772660 (9.3 MiB)  TX bytes:280592087 (267.5 MiB)

After when I check on "Console" on the Host, I have all time this error on the disconnect:

13:22:43.454970 -0500	lsd	LSExceptions shared instance invalidated for timeout.
13:22:43.479991 -0500	kernel	X86PlatformPlugin result 0
13:22:43.479997 -0500	kernel	X86PlatformPlugin result 0
13:22:43.743794 -0500	locationd	{"msg":"kCLConnectionMessageWatchdog", "event":"activity", "this":"0x7fdcf0c808d0", "registrationRequired":0, "registrationReceived":1}
13:22:43.776090 -0500	kernel	vboxnet0 duplicate IP address 172.21.21.1 sent from address 08:00:27:5a:96:1c

After I installed Wireshark to inspect the netwok and you can view the ARP package from this MAC:

I open this ticket because this mac address start with "08:00:27:5a" look to be the DHCPServer assignments.

Can you please have a check on this issue ?

[TB-Quentin]

The DHCP servers and Host Only Interfaces were set with the same IP address. This was discovered to cause a problem on the guest because the Virtual Box engine actually gives both of those subsystems a different virtual MAC.

substance@work:~$ sudo arping 172.21.21.1
ARPING 172.21.21.1
42 bytes from 0a:00:27:00:00:00 (172.21.21.1): index=0 time=300.050 usec
58 bytes from 08:00:27:07:66:7e (172.21.21.1): index=1 time=350.147 usec
42 bytes from 0a:00:27:00:00:00 (172.21.21.1): index=2 time=203.526 usec
58 bytes from 08:00:27:07:66:7e (172.21.21.1): index=3 time=231.830 usec

So the symptom on the guest is intermittent ARP poisoning, leading to double responses to an arping from the guest to the host only interface ip address.

WORK:

substance@work:~$ arp -a
? (10.0.2.2) at 52:54:00:12:35:02 [ether]  on eth0
? (172.21.21.1) at 0a:00:27:00:00:00 [ether]  on eth1  <====
? (172.18.0.7) at 02:42:ac:12:00:07 [ether]  on br-2e620a16c6b5

FAIL:

substance@work:~$ arp -a
? (10.0.2.2) at 52:54:00:12:35:02 [ether]  on eth0
? (172.21.21.1) at 08:00:27:07:66:7e [ether]  on eth1  <====
? (172.18.0.7) at 02:42:ac:12:00:07 [ether]  on br-2e620a16c6b5

Same error here: ​https://forums.virtualbox.org/viewtopic.php?f=7&t=58761

[Valery Ushakov]

Please, can you do ps -ef | grep vboxnet0 and attach/paste the config file for the dhcp server. Probably ~/Library/VirtualBox/HostInterfaceNetworking-vboxnet0-Dhcpd.config