On Tuesday 18 Oct 2005 16:21, Greig A Cowan wrote:
> Hi Graeme,
>
> I upgraded to 1.3.8 yesterday and have been seeing the same error messages
> in our DPM (log at end), although I haven't seem the large time delays
> that you are experiencing.
OK resolved that one. It seemed to be a hangover from ui1-gla not being quite
sure whether se2-gla was an internal or external hosts. Panic over!
>
> I am still having issues with my migration of our Classic SE files into
> our DPM (converting the classic SE into a pool node). As an update to my
> email from yesterday, it now appears that although the migration created a
> whole set of /dpm/epcc.ed.ac.uk/home/dteam/generated/YYYY-MM-DD
> directories and populated them with files, they don't match to the files
> in the corresponding classic SE directories (the names are different and
> not all of the YYYY-MM-DD dirs are present in the DPNS). I'm confused :|
I haven't tried this - and actually we're unlikely to, because our classic SE
is looking at an nfs mounted partition on another host. We'd rather that all
went away, rather than being migrated ;-)
However, from a brief perusal of the migration script, it looks like it tries
to add NS replicas for each file on the classic SE, so that any file on the
classic with path
classic-se:/store/VO/PATH/FILE
gets added to the DPNS as
/dpm/DOMAIN/home/VO/PATH/FILE
If that's not happened (you indicate it didn't) then I think something went
wrong with the script. Did you get errors (from l158-159 of migration.c)?
Might be worth looking at the Cns_file_replica table in the DPNS.
Sorry I can't be more helpful this side of running the damn thing.
g
--
--------------------------------------------------------------------
Dr Graeme Stewart http://www.physics.gla.ac.uk/~graeme/
GridPP DM Wiki http://wiki.gridpp.ac.uk/wiki/Data_Management
|