Lee;
I pushed a fix to the nglims fork on bitbucket which should resolve this
errors. Apologies, the nglims error is a mistake on my part. The way
Galaxy handles version updates changed a bit on the infrastructure side
and I hadn't coded correctly to handle all the cases.

You shouldn't need to restart and lose all your data. I think you're
nearly there with the upgrade. More details below and thanks for all the
patience.

> 114 -> 115...
[...]

> OperationalError: (OperationalError) there is already another table or
> index with this name: migration_tmp u'ALTER TABLE galaxy_user R
>
>               ENAME TO migration_tmp' ()

This is the same error as before, so needs the same hack to work around:

$ sqlite3 database/universe.sqlite
sqlite> drop table migration_tmp;

Sorry about that one, not sure exactly how to fix this more cleanly.

> 115 -> 116...
>
> Tables for a next generation sequencing lims tied to requests.

[...]

> sqlalchemy.exc.InvalidRequestError: Table 'sample_request_map' is already
> defined for this MetaData instance.  Specify 'extend_existin
>
>             g=True' to redefine options and columns on an existing Table
> object.

This is my bug mentioned above, so should work with an updated checkout
of the code.

Hope this gets you there. Fingers crossed,
Brad