| From: | David Wall <d(dot)wall(at)computer(dot)org> | 
|---|---|
| To: | pgsql-jdbc(at)postgresql(dot)org | 
| Subject: | JDBC with PG 9.2 and large object permissions as they relate to blobs (after upgrade from 8.4) | 
| Date: | 2012-12-17 18:31:56 | 
| Message-ID: | 50CF651C.9040004@computer.org | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-jdbc | 
We upgraded our PG from 8.4 to 9.2 and found that our pg_dump and 
pg_restore caused all of the large objects to be given permission to our 
db admin user, but not our application user.  This resulted in some null 
pointers in our java code when trying to access an OID column, with the 
PG error saying: permission denied for large object NNNN.
I checked with the excellent PG team and they pointed out that prior to 
9.0 large objects didn't have any access permissions/roles assigned to 
them.  So I guess on restore, all of the large objects became owned by 
our DB admin account, and despite all of our application role table 
GRANTS, these permission changes did not go down to the related large 
objects.
I'm just checking if the latest JDBC library is doing anything special 
with respect to large objects and permissions.  We just use the 
PreparedStatement getBlob()/setBlob().  Will the large objects be owned 
by the role that executes the INSERT?
By the way, we used the following psql script (found in posting 
http://doginpool.blogspot.com/2011/10/today-upgraded-to-new-ubuntu-11.html) 
to give all large objects ownership back to our application role and 
this seems to have resolved our issue:
do $$
declare r record;
begin
for r in select loid from pg_catalog.pg_largeobject loop
execute 'ALTER LARGE OBJECT ' || r.loid || ' OWNER TO DBUSER';
end loop;
end$$;
CLOSE ALL;
Thanks,
David
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Dave Cramer | 2012-12-17 18:44:43 | Re: JDBC with PG 9.2 and large object permissions as they relate to blobs (after upgrade from 8.4) | 
| Previous Message | Scott Harrington | 2012-12-15 16:19:07 | Re: performance problem of Failover Datasource? |