"Unable to open virtual bridge" Host: linux Guest: XP

Discussion in 'Windows Guest OS Discussion' started by tellmejeff, Mar 29, 2007.

  1. tellmejeff

    tellmejeff Bit poster

    Messages:
    4
    Periodically I get the error about being unable to open the virtual bridge when running XP as a guest OS under linux. I haven't yet found a clear solution to this problem. Seems like if I reboot it sometimes goes away, but not always. I don't even really understand what the problem is. My ethernet connection is fine and I can bring up the virtual nic just fine. Is there some error log that I should check that might give me a better idea of what is wrong? Has someone dealt with this and can just tell me how I fix it when it happens. It is kind of the most annoying part of Parallels right now. Also, it doesn't seem to happen when I use other OS's under Parallels (like another flavor of linux). It seems odd that XP would be a culprit, but somehow it wouldn't completely surprise me.

    -jp
     
  2. ms @ parallels

    ms @ parallels Parallels Team

    Messages:
    87
    When you get this problem again, please post output of the following commands to this thread:

    1. dmesg
    2. lsmod
    2. ps aux | grep -i prl

    thx.
     
  3. tellmejeff

    tellmejeff Bit poster

    Messages:
    4
    * I didn't see any relevant ouput in dmesg (nothing added when error appears).
    * Mandriva doesn't have lsmod
    * I do see the dhcpd daemon running for the vnic0 if that is what you expected:

    bash> ps -edf | grep prl
    root 18393 1 0 10:24 ? 00:00:00 /usr/lib/parallels/autostart/prl_dhcpd -d vnic0
     
  4. tellmejeff

    tellmejeff Bit poster

    Messages:
    4
    still getting "unable to open virtual bridge", host linux, guest os XP

    This is still happening and its really annoying. Here is the output from dmesg:

    [ 8297] parallels Hypervisor 2.2-2112 exited.
    [ 8383] Device /dev/hypervisor registered with 10.60
    [ 8383] parallels Hypervisor 2.2-2112 initialized.
    [ 8397] Device /dev/vm-main registered with 10.59
    [ 8397] Parallels Workstation 2.2-2112 Main module initialized.
    device vnic0 entered promiscuous mode
    audit(1181581343.167:6): dev=vnic0 prom=256 old_prom=0 auid=4294967295
    Floppy drive(s): fd0 is 1.44M
    floppy0: no floppy controllers found
    vnic0: no IPv6 routers present

    I don't have lsmod. Here is the ps output you wanted:

    xfs 3600 0.0 0.1 5180 3128 ? Ss 12:36 0:01 xfs -port -1 -daemon -droppriv -user xfs
    jpoore 6590 0.0 0.0 2852 932 ? Ss 12:37 0:00 /usr/bin/dbus-daemon --fork --print-pid 9 --print-address 7 --session
    jpoore 8625 0.0 0.0 2912 808 pts/3 R+ 13:07 0:00 grep --color -i pri
     
  5. electricbob

    electricbob Bit poster

    Messages:
    1
    I'm seeing the same problem with the latest update on Mac OS X as a host, XP as the hosted OS. Haven't seen this problem before on Mac OS X.

    Thanks! - Bob
     
  6. titetanium

    titetanium Member

    Messages:
    84
    While I haven't gotten this error, I do have a blown vnic0 interface. I re-ran parallels-config to rebuild the drivers but -- the parallels dhcp server refuses to load. My configuration is set to use host-only networking and I did the necessary steps to enable host-only networking to work (it can access both the host computer and the internet).

    This thread said something about using the same ip address as the vnic0 dhcp server will cause it to be taken out of the game, I think it might relevant to the OP's thread.

    Upon starting the host computer, I saw a line in the bootup screen where firestarter failed to start because vnic0 wasn't available for it to properly initialize and masquerade upon startup. Ifconfig does not show vnic0 as being up. I can bring it up manually with
    Code:
    ifconfig vnic0 10.37.129.2 netmask 255.255.255.0
    and everything works fine after that. I created a script in /etc/init.d called vnic0 to bring up the vnic0 interface before the firewall loads, but I feel this is a kludgy hack, not an elegant permanent solution.

    Here is the relevant bits if its a clue:
    dmesg:
    device vnic0 entered promiscuous mode
    audit(1190555774.735:4): dev=vnic0 prom=256 old_prom=0 auid=4294967295
    [ 9282] Can't enable interrupts via /dev/rtc: -25
    (/usr/lib/parallels/drivers/drv_main/common/timer.c:176)
    device vnic0 left promiscuous mode
    audit(1190555992.239:5): dev=vnic0 prom=0 old_prom=256 auid=4294967295

    lsmod:
    Module Size Used by
    vmvirtualnic 2116 0
    vm_bridge 10524 0
    vm_main 37488 0
    hypervisor 73128 1 vm_main

    Any help?
     
  7. titetanium

    titetanium Member

    Messages:
    84
    Turns out that uninstalling parallels, renaming the /usr/lib/parallels to a backup folder, reinstalling parallels, and finally re-compiling the parallels modules reinitialized the broken dhcp server component in the vnic0 driver.

    However, if the parallels dhcp server is turned off by means of dhcp=0, then vnic0 will not start but it can bring it up manually though.

    HTH,
    Titetanium
     

Share This Page