[tahoe-dev] Unhappiness Error

[Brad Rupp]

On 9/27/2012 12:21 PM, David-Sarah Hopwood wrote:
> On 27/09/12 17:27, Brad Rupp wrote:
>> I am getting a strange error while uploading files to Tahoe today.  The error is:
>>
>> "UploadUnhappinessError('shares could be placed on only 9 server(s) such that any 7 of
>> them have enough shares to recover the file, but we were asked to place shares on at least
>> 10 such servers. (placed all 12 shares, want to place shares on at least 10 servers such
>> that any 7 of them have enough shares to recover the file, sent 10 queries to 10 servers,
>> 10 queries placed some shares, 0 placed none (of which 0 placed none due to the server
>> being full and 0 placed none due to an error))',) (profiles_activity.frm)"
>>
>> My client settings are:
>>
>> shares.needed = 7
>> shares.happy = 10
>> shares.total = 12
>>
>> Actual available storage servers: 16
>>
>> What is strange about the error is that it shows that 10 shares were placed and that there
>> were 0 errors.  However, I get an unhappiness error.
>>
>> The part of the error message that has me confused is:
>>
>> "sent 10 queries to 10 servers, 10 queries placed some shares, 0 placed none (of which 0
>> placed none due to the server being full and 0 placed none due to an error)"
>>
>> What am I missing?
>
> That seems like a bug. It can happen that the happiness achieved is less than the number
> of shares placed, but in that case the error message should be showing that some servers
> didn't accept shares in the first-choice placement.
>
> If the servers are still running, can you check their foolscap logs for any errors around
> the time of the upload? Briefly, that is:
>
> On each storage server,
>   - install foolscap, e.g. using 'sudo apt-get install python-foolscap' on
>     Debian-based Linux
>   - run 'bin/tahoe @flogtool tail --catch-up STORAGE_SERVER_BASEDIR/private/logport.furl'
>
> See docs/logging.rst has more information about logging.
>
> Is it Tahoe-LAFS 1.9.2 running on the client and all storage servers?
>

This is happening on the VolunteerGrid2, so I only have access to two of 
the storage servers.  The remaining storage servers are owned by others. 
  I could try to get the logs, but that is going to take some serious time.

My two nodes and client are running version 1.9.2.  There is a mix of 
version 1.8.3, 1.9.1 and 1.9.2 storage servers in the grid.

Brad