#69 Segfault after 3 months working fine

v1.0 (example)
open
nobody
2
2014-01-16
2013-12-05
No

Yesterday I experienced a segfault on my vde_switch after 88 days working fine. VDE connects my qemu VMs together and is heavily in use. I want to report this to help you find problems and to ask if there is a possibility to restart the vde device after this fault and how this will affect my VMs.

I use the debian package (vde2, 2.3.2-4, amd64, Linux 3.2.0-4-amd64 #1 SMP Debian 3.2.46-1 x86_64 GNU/Linux), the log is as follows:
Dec 4 22:59:46 server kernel: [7847125.283822] vde_switch[2338]: segfault at 80 ip 0000000000404b48 sp 00007fffbd1339e8 error 4 in vde_switch[400000+15000]

Thank you and VDE is great!
mtmayr

Related

Bugs: #69

Discussion

  • renzo davoli
    renzo davoli
    2013-12-08

    On Thu, Dec 05, 2013 at 03:29:19PM +0000, Mathias Mayrhofer wrote:

    [bugs:#69] Segfault after 3 months working fine

    Status: open
    Labels: segfault vde_switch crash
    Created: Thu Dec 05, 2013 03:29 PM UTC by Mathias Mayrhofer
    Last Updated: Thu Dec 05, 2013 03:29 PM UTC
    Owner: nobody

    Yesterday I experienced a segfault on my vde_switch after 88 days working fine. VDE connects my qemu VMs together and is heavily in use. I want to report this to help you find problems and to ask if there is a possibility to restart the vde device after this fault and how this will affect my VMs.

    I use the debian package (vde2, 2.3.2-4, amd64, Linux 3.2.0-4-amd64 #1 SMP Debian 3.2.46-1 x86_64 GNU/Linux), the log is as follows:
    Dec 4 22:59:46 server kernel: [7847125.283822] vde_switch[2338]: segfault at 80 ip 0000000000404b48 sp 00007fffbd1339e8 error 4 in vde_switch[400000+15000]

    Thank you and VDE is great!
    Thank you for your kind report.

    I'll try to see what I can do to fix the bug. Have you any other symptom to report? Hunting for a bug using just the
    instruction pointer can be a bit daunting...

    It is not currently possible for VM to be automatically reconnected to a new switch.
    Automatic reconnection can be a useful feature to study for a future versions of libvdeplug.
    Unfortunately, the only current way to have your VMs reconnected is to restart them.

    renzo davoli
    
     

    Related

    Bugs: #69

  • Hi Renzo Davoli!

    I cannot report any other symptoms, because it worked fine until the segfault. The network device was still up due ifconfig, but the process vde_switch went away. Below you see the ps aux right before the segfault and ifconfig after the segfault and how it is referred by with qemu. There were no other log entries nearby the fault.

    Thanks for explaining, greetings!
    mtmayr

    F S UID        PID  PPID  C PRI  NI ADDR SZ WCHAN  STIME TTY          TIME CMD
    5 S root      2338     1  0  80   0 -  3698 -      Sep05 ?        11:11:02 /usr/bin/vde_switch -s /var/run/vde2/vmn.ctl -m 660 -g vde2-net -p /var/run/vde2/vmn.pid -t vmn -M /var/run/vde2/vmn.mgmt --mgmtmode 660 -d -t vmn
    
    vmn       Link encap:Ethernet  HWaddr 12:74:99:ce:14:3f
              inet addr:192.168.1.1  Bcast:192.168.1.255  Mask:255.255.255.0
              UP BROADCAST MULTICAST  MTU:1500  Metric:1
              RX packets:471448132 errors:0 dropped:0 overruns:0 frame:0
              TX packets:354058105 errors:0 dropped:0 overruns:0 carrier:0
              collisions:0 txqueuelen:500
              RX bytes:176797413967 (164.6 GiB)  TX bytes:234534002363 (218.4 GiB)
    
    qemu -net vde,sock=/var/run/vde2/vmn.ctl [...]
    
    
     
    Last edit: Mathias Mayrhofer 2013-12-10