[sheepdog-users] Force cluster to start with a missing node

[Liu Yuan]

On 01/12/2013 03:22 PM, Valerio Pachera wrote:
> 2012/12/21 Valerio Pachera <sirio81 at gmail.com>:
>> we stop the cluster for do some maintenance.
>> We find out that a node has hardware problems and we do not want (or
>> can't) use it again.
> 
> I tried to simulate this situation.
> Clister with 4 node, formated with 3 copies and quorum.
> Collie cluster shutdown,
> run sheep on 3 nodes,
> collie cluster recover force
> 
> The cluster startds but I have a node that shows only one vdi:
> 
> root at sheepdog001:~# collie vdi list
>   Name        Id    Size    Used  Shared    Creation time   VDI id  Copies  Tag
>   win7         1   30 GB   30 GB  0.0 MB 2013-01-08 19:07   605eca
> 3
>   squeeze      1  1.2 GB  1.2 GB  0.0 MB 2013-01-08 18:53   a34c67
> 3
>   test2        1  4.0 GB  1.7 GB  0.0 MB 2013-01-08 21:07   fd3815     2
> 
> root at sheepdog002:~# collie vdi list
>   Name        Id    Size    Used  Shared    Creation time   VDI id  Copies  Tag
>   win7         1   30 GB   30 GB  0.0 MB 2013-01-08 19:07   605eca     3
> 
> version 0.5.5_36_g6101dbe
> 

How did you create the vdi? It seems that object cache isn't flushed.
you can use 'vdi cache flush xxx' to assure vdi is flushed.

Thanks,
Yuan