[sheepdog] [PATCH v1] sheep/md: leave cluster if disk number equals zero in handling EIO

[Valerio Pachera]

I confirm such wrong behaviour in Sheepdog daemon version 0.8.0_331_gecc533e.
More than leaving the cluster, the node should get weight 0 and become
a gateway only.
This way guests can continue working.

root at test007:~# dog node list
  Id   Host:Port         V-Nodes       Zone
   0   192.168.10.4:7000        126   67807424
   1   192.168.10.5:7000        129   84584640
   2   192.168.10.6:7000        129  101361856
   3   192.168.10.7:7000        129  118139072

root at test007:~# dog node md info --all
Id      Size    Used    Avail   Use%    Path
Node 0:
 0      216 GB  5.7 GB  210 GB    2%    /mnt/sheep/0
Node 1:
 0      219 GB  4.8 GB  215 GB    2%    /mnt/sheep/0
Node 2:
 0      220 GB  5.2 GB  215 GB    2%    /mnt/sheep/0
Node 3:
 0      220 GB  5.2 GB  215 GB    2%    /mnt/sheep/0

root at test007:~# dog node md unplug /mnt/sheep/0
    __
   ()'`;
   /\|`
  /  |   Caution! Recovery will be triggered after unplugging.
(/_)_|_  Are you sure you want to continue? [yes/no]: yes

root at test007:~# dog node list
  Id   Host:Port         V-Nodes       Zone
   0   192.168.10.4:7000        126   67807424
   1   192.168.10.5:7000        129   84584640
   2   192.168.10.6:7000        129  101361856
   3   192.168.10.7:7000        129  118139072
root at test007:~# dog node recovery
Nodes In Recovery:
  Id   Host:Port         V-Nodes       Zone       Progress
   3   192.168.10.7:7000     129  118139072        0.0%

root at test007:~# dog node info
Id      Size    Used    Avail   Use%
 0      216 GB  5.7 GB  210 GB    2%
 1      219 GB  4.8 GB  215 GB    2%
 2      220 GB  5.2 GB  215 GB    2%
 3      0.0 MB  0.0 MB  0.0 MB    0%
(It hangs...I had to press ctrl+c)