[sheepdog-users] Something happens when killing a guest

[Hitoshi Mitake]

Thanks for your report, I created an issue:
https://github.com/sheepdog/sheepdog/issues/164

The problem seems to be caused by checkpoint syncing. The feature is
required by VDI locking. If we turn off the feature (format cluster
without -l), the sync isn't required. I'll insert a conditional branch
for doing the checkpoint sync based on the -l option.

Thanks,
Hitoshi


On Fri, Sep 18, 2015 at 12:19 AM, Valerio Pachera <sirio81 at gmail.com> wrote:
> This is my cluster
>
> dog node list
>   Id   Host:Port         V-Nodes       Zone
>    0   192.168.6.45:7000        644  755411136
>    1   192.168.6.47:7000        324  788965568
>    2   192.168.6.48:7000        432  805742784
>
> Sheepdog daemon version 0.9.0_278_g6ca1f21
>
> I was running a single guest.
> I tried to kill it.
> For few seconds ssh doesn't respond.
> In sheep.log I see that:
>
> Sep 17 17:14:39   WARN [block] do_add_vdi_state(379) caution, updating
> parent VID of VID: a34a06 (old parent VID: a34a05, new parent VID:
> a34a05)
> Sep 17 17:14:39   INFO [main] local_vdi_state_checkpoint_ctl(1479)
> getting vdi state checkpoint at epoch 0
> Sep 17 17:14:39   WARN [block] do_add_vdi_state(379) caution, updating
> parent VID of VID: a34a06 (old parent VID: a34a05, new parent VID:
> a34a05)
> Sep 17 17:14:39   INFO [main] local_vdi_state_checkpoint_ctl(1479)
> getting vdi state checkpoint at epoch 0
> Sep 17 17:14:39   INFO [main] local_vdi_state_checkpoint_ctl(1479)
> getting vdi state checkpoint at epoch 0
> Sep 17 17:14:39   INFO [main] local_vdi_state_checkpoint_ctl(1479)
> getting vdi state checkpoint at epoch 0
> Sep 17 17:14:39   INFO [main] local_vdi_state_checkpoint_ctl(1479)
> freeing vdi state checkpoint at epoch 0
> Sep 17 17:14:39   INFO [main] local_vdi_state_checkpoint_ctl(1479)
> getting vdi state checkpoint at epoch 0
> Sep 17 17:14:39   INFO [main] get_vdi_state_checkpoint(2178) get
> request for not prepared vdi state checkpoint, epoch: 0
> Sep 17 17:14:39   INFO [main] local_vdi_state_checkpoint_ctl(1489)
> failed to get vdi state checkpoint: Ask to try again
> Sep 17 17:14:39   INFO [main] recover_object_main(1002) object
> recovery progress   1%
> Sep 17 17:14:39   INFO [main] recover_object_main(1002) object
> recovery progress   2%
> ...<cut>...
> --
> sheepdog-users mailing lists
> sheepdog-users at lists.wpkg.org
> https://lists.wpkg.org/mailman/listinfo/sheepdog-users