[Xen-users] Can't destroy zombie domain, holding onto resources

[Andy Smith <andy@xxxxxxxxxxxxxx>]

Hi,

On restarting one of my domains, it went into zombie state and
failed to restart.  Now I see it like this:

Name                              ID Mem(MiB) VCPUs State  Time(s)
Domain-0                           0      256     1 r----- 1088916.2
Zombie-test1                     111      128     1 ---s-d 195171.2

Issuing "xm shutdown 111" or "xm destroy 111" immediately returns to
command prompt with no action seemingly being taken.

The "test1" domain cannot be started again, giving the "hotplug
scripts not working" error.

LVm believes that the domain's root and swap block devices are still
in use.  "ifconfig" and "ip link" show the domain's vif as still
being present, though it is not currently part of xenbr0.

I can see multiple instances of the domain's vbds and vif if I use
xenstore-ls:

$ sudo xenstore-ls | less
[...]
local = ""
 domain = ""
[...]
     111 = ""
      2049 = ""
       domain = "test1"
       frontend = "/local/domain/111/device/vbd/2049"
       dev = "sda1"
       state = "4"
       params = "mainvg/domu_test1_root"
       mode = "w"
       frontend-id = "111"
       type = "phy"
       physical-device = "fd:15"
       hotplug-status = "connected"
       sectors = "31457280"
       info = "0"
       sector-size = "512"
      2050 = ""
       domain = "test1"
       frontend = "/local/domain/111/device/vbd/2050"
       dev = "sda2"
       state = "4"
       params = "mainvg/domu_test1_swap"
       mode = "w"
       frontend-id = "111"
       type = "phy"
       physical-device = "fd:14"
       hotplug-status = "connected"
       sectors = "524288"
       info = "0"
       sector-size = "512"
     160 = ""
      2049 = ""
       domain = "test1"
       frontend = "/local/domain/160/device/vbd/2049"
       dev = "sda1"
       state = "1"
       params = "mainvg/domu_test1_root"
       mode = "w"
       frontend-id = "160"
       type = "phy"
      2050 = ""
       domain = "test1"
       frontend = "/local/domain/160/device/vbd/2050"
       dev = "sda2"
       state = "1"
       params = "mainvg/domu_test1_swap"
       mode = "w"
       frontend-id = "160"
       type = "phy"
     161 = ""
      2049 = ""
       domain = "test1"
       frontend = "/local/domain/161/device/vbd/2049"
       dev = "sda1"
       state = "1"
       params = "mainvg/domu_test1_root"
       mode = "w"
       frontend-id = "161"
       type = "phy"
      2050 = ""
       domain = "test1"
       frontend = "/local/domain/161/device/vbd/2050"
       dev = "sda2"
       state = "1"
       params = "mainvg/domu_test1_swap"
       mode = "w"
       frontend-id = "161"
       type = "phy"
     162 = ""
      2049 = ""
       domain = "test1"
       frontend = "/local/domain/162/device/vbd/2049"
       dev = "sda1"
       state = "1"
       params = "mainvg/domu_test1_root"
       mode = "w"
       frontend-id = "162"
       type = "phy"
      2050 = ""
       domain = "test1"
       frontend = "/local/domain/162/device/vbd/2050"
       dev = "sda2"
       state = "1"
       params = "mainvg/domu_test1_swap"
       mode = "w"
       frontend-id = "162"
       type = "phy"
     163 = ""
      2049 = ""
       domain = "test1"
       frontend = "/local/domain/163/device/vbd/2049"
       dev = "sda1"
       state = "1"
       params = "mainvg/domu_test1_root"
       mode = "w"
       frontend-id = "163"
       type = "phy"
      2050 = ""
       domain = "test1"
       frontend = "/local/domain/163/device/vbd/2050"
       dev = "sda2"
       state = "1"
       params = "mainvg/domu_test1_swap"
       mode = "w"
       frontend-id = "163"
       type = "phy"
     164 = ""
      2049 = ""
       domain = "test1"
       frontend = "/local/domain/164/device/vbd/2049"
       dev = "sda1"
       state = "1"
       params = "mainvg/domu_test1_root"
       mode = "w"
       frontend-id = "164"
       type = "phy"
      2050 = ""
       domain = "test1"
       frontend = "/local/domain/164/device/vbd/2050"
       dev = "sda2"
       state = "1"
       params = "mainvg/domu_test1_swap"
       mode = "w"
       frontend-id = "164"
       type = "phy"
[...]
    vif = ""
[...]
     111 = ""
      0 = ""
       bridge = "xenbr0"
       domain = "test1"
       handle = "0"
       script = "/etc/xen/scripts/vif-bridge"
       state = "4"
       frontend = "/local/domain/111/device/vif/0"
       mac = "00:16:3e:00:67:fc"
       vifname = "v-test1"
       frontend-id = "111"
       hotplug-status = "connected"
     160 = ""
      0 = ""
       bridge = "xenbr0"
       domain = "test1"
       handle = "0"
       script = "/etc/xen/scripts/vif-bridge"
       state = "5"
       frontend = "/local/domain/160/device/vif/0"
       mac = "00:16:3e:00:67:fc"
       vifname = "v-test1"
       frontend-id = "160"
     161 = ""
      0 = ""
       bridge = "xenbr0"
       domain = "test1"
       handle = "0"
       script = "/etc/xen/scripts/vif-bridge"
       state = "5"
       frontend = "/local/domain/161/device/vif/0"
       mac = "00:16:3e:00:67:fc"
       vifname = "v-test1"
       frontend-id = "161"
     162 = ""
      0 = ""
       bridge = "xenbr0"
       domain = "test1"
       handle = "0"
       script = "/etc/xen/scripts/vif-bridge"
       state = "5"
       frontend = "/local/domain/162/device/vif/0"
       mac = "00:16:3e:00:67:fc"
       vifname = "v-test1"
       frontend-id = "162"
     163 = ""
      0 = ""
       bridge = "xenbr0"
       domain = "test1"
       handle = "0"
       script = "/etc/xen/scripts/vif-bridge"
       state = "5"
       frontend = "/local/domain/163/device/vif/0"
       mac = "00:16:3e:00:67:fc"
       vifname = "v-test1"
       frontend-id = "163"
     164 = ""
      0 = ""
       bridge = "xenbr0"
       domain = "test1"
       handle = "0"
       script = "/etc/xen/scripts/vif-bridge"
       state = "5"
       frontend = "/local/domain/164/device/vif/0"
       mac = "00:16:3e:00:67:fc"
       vifname = "v-test1"
       frontend-id = "164"

This is xen 3.0.2-3 on debian sarge from sarge backports.

Anyone have any ideas how to destroy this domain and free up the
resources it has a hold on, without rebooting dom0?

Cheers,
Andy

Attachment: signature.asc
Description: Digital signature

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users