host-only networking slows down, high latency

[Scott Fehrman]

MacOS 10.5 using VBox 2.2.2 with OpenSolaris guest. Started having networking issues after upgrading to 2.2.2. Image works fine for 1-2 hours, using ssh / scp from host to guest. Then for no obvious reason the networking has large delays (2-10 seconds), type in command in ssh terminal and wait. Here is some "ping" data ...

macbook:[sfehrman] ping sedemo8
PING sedemo (192.168.100.101): 56 data bytes
64 bytes from 192.168.100.101: icmp_seq=0 ttl=255 time=13851.959 ms
64 bytes from 192.168.100.101: icmp_seq=1 ttl=255 time=12851.360 ms
64 bytes from 192.168.100.101: icmp_seq=2 ttl=255 time=11850.678 ms
64 bytes from 192.168.100.101: icmp_seq=3 ttl=255 time=10850.062 ms
64 bytes from 192.168.100.101: icmp_seq=4 ttl=255 time=9849.439 ms
64 bytes from 192.168.100.101: icmp_seq=5 ttl=255 time=8848.794 ms
64 bytes from 192.168.100.101: icmp_seq=6 ttl=255 time=7848.265 ms
64 bytes from 192.168.100.101: icmp_seq=7 ttl=255 time=6847.562 ms
64 bytes from 192.168.100.101: icmp_seq=8 ttl=255 time=5846.956 ms
64 bytes from 192.168.100.101: icmp_seq=9 ttl=255 time=4846.243 ms
64 bytes from 192.168.100.101: icmp_seq=10 ttl=255 time=3845.652 ms
64 bytes from 192.168.100.101: icmp_seq=11 ttl=255 time=2845.027 ms
64 bytes from 192.168.100.101: icmp_seq=12 ttl=255 time=1844.426 ms
64 bytes from 192.168.100.101: icmp_seq=13 ttl=255 time=843.721 ms
64 bytes from 192.168.100.101: icmp_seq=14 ttl=255 time=29428.120 ms
64 bytes from 192.168.100.101: icmp_seq=15 ttl=255 time=28427.485 ms
64 bytes from 192.168.100.101: icmp_seq=16 ttl=255 time=27426.888 ms
64 bytes from 192.168.100.101: icmp_seq=17 ttl=255 time=26426.605 ms
64 bytes from 192.168.100.101: icmp_seq=18 ttl=255 time=25426.004 ms
64 bytes from 192.168.100.101: icmp_seq=19 ttl=255 time=24425.402 ms
64 bytes from 192.168.100.101: icmp_seq=20 ttl=255 time=23424.780 ms
64 bytes from 192.168.100.101: icmp_seq=21 ttl=255 time=22424.066 ms
64 bytes from 192.168.100.101: icmp_seq=22 ttl=255 time=21423.472 ms
64 bytes from 192.168.100.101: icmp_seq=23 ttl=255 time=20421.088 ms
64 bytes from 192.168.100.101: icmp_seq=24 ttl=255 time=19420.491 ms
64 bytes from 192.168.100.101: icmp_seq=25 ttl=255 time=18419.905 ms
64 bytes from 192.168.100.101: icmp_seq=26 ttl=255 time=17419.179 ms
64 bytes from 192.168.100.101: icmp_seq=27 ttl=255 time=16418.473 ms
64 bytes from 192.168.100.101: icmp_seq=28 ttl=255 time=15417.836 ms
64 bytes from 192.168.100.101: icmp_seq=29 ttl=255 time=14417.111 ms
64 bytes from 192.168.100.101: icmp_seq=30 ttl=255 time=13416.476 ms
64 bytes from 192.168.100.101: icmp_seq=31 ttl=255 time=12415.752 ms
64 bytes from 192.168.100.101: icmp_seq=32 ttl=255 time=11415.001 ms
64 bytes from 192.168.100.101: icmp_seq=33 ttl=255 time=10414.306 ms
64 bytes from 192.168.100.101: icmp_seq=34 ttl=255 time=9413.643 ms
64 bytes from 192.168.100.101: icmp_seq=35 ttl=255 time=8842.514 ms
64 bytes from 192.168.100.101: icmp_seq=36 ttl=255 time=7841.669 ms
64 bytes from 192.168.100.101: icmp_seq=37 ttl=255 time=6839.962 ms
64 bytes from 192.168.100.101: icmp_seq=38 ttl=255 time=5839.377 ms
64 bytes from 192.168.100.101: icmp_seq=39 ttl=255 time=4837.303 ms
64 bytes from 192.168.100.101: icmp_seq=40 ttl=255 time=3836.596 ms
64 bytes from 192.168.100.101: icmp_seq=41 ttl=255 time=2835.894 ms
64 bytes from 192.168.100.101: icmp_seq=42 ttl=255 time=1835.271 ms
64 bytes from 192.168.100.101: icmp_seq=43 ttl=255 time=834.580 ms
--- sedemo ping statistics ---
71 packets transmitted, 44 packets received, 38% packet loss
round-trip min/avg/max/stddev = 834.580/12596.713/29428.120/8208.305 ms

A reboot of the image will restore normal host-to-guest performance, for another few hours.

[Scott Fehrman]

vbox log file