Looking on IBM's site for HY009 seems to point to a C code errror ... which, if this is all you're doing, apparently straight out of the DBD::DB2 docs, doesn't seem like a reasonable return code for that operation. So I'd suggest opening a PMR with IBM.

Personally, I've always just cataloged the remote db's locally - it makes for easier debugging when I can just run quick command-line queries to see what I'm doing wrong. I also like how I can have an instance that catalogs a test database, and just recatalog the database alias to point at the production instance when I'm done testing. ;-)