[sheepdog-users] failed to read inode header

[Maxim Terletskiy]

Hi!

After some inspecting I found that all(I hope all) missing data pieces 
left on one disk which I forgot to plug after sheep restart.

Can I just plug this disk to working sheep to access this data? Or I 
need to stop sheep daemon and run it again with this disk in options? 
Maybe there is some another safe method to return this data to cluster?

Sorry, if I'm asking too much questions. :) This data is extremely 
important for me and unfortunately I don't have any backups. :(

29.05.2014 16:29, Maxim Terletskiy пишет:
> Hi!
>
> After nodes recovery I'm running "dog vdi check" on some images for 
> test. "Check" reporting about missing pieces of data and finally say 
> "finish check&repair".
>
> I'm trying to test image by "qemu-img convert" but it is failing. 
> Doing "dog vdi check" again and see that data pieces can't be found. I 
> see pieces of data with this names on previously failed node in .stale 
> directory. Can I recover this pieces manually?
>
> 29.05.2014 12:32, Maxim Terletskiy пишет:
>> Running 0.8.0.
>>
>> Deleted all files by mask 80* from data dirs and after that sheep 
>> daemon started successfully and joined cluster. After that I'm able 
>> to run any command which failed before with message about "failed to 
>> connect" to failed node.
>>
>> Now waiting for node to end check.
>>
>> By the way how can I assume that my data is consistent and no pieces 
>> were lost?
>>
>> 29.05.2014 9:52, Liu Yuan пишет:
>>> On Wed, May 28, 2014 at 07:12:39PM +0400, Maxim Terletskiy wrote:
>>>> HI!
>>>>
>>>> Now I see next things. 3 of 4 nodes online, they are replicating as
>>>> I see via iostat and vnstat. Fourth node can't start sheep daemon.
>>>> Dying with this error:
>>>> May 28 18:04:42 ERROR [main] init_vdi_state(213) failed to read
>>>> inode header 80bb794400000000 0
>>>>
>>> which version do you run? As far as I know, this problem was fixed 
>>> long ago
>>>
>>> Thanks
>>> Yuan
>>
>