We have a new QuickFinder 5.0 installation from OES2 that we're trying to use to index a large (700G) NSS volume (served via OES2/Linux via NCP). The indexing engine starts the indexing, but after a period of time (it's varied anywhere from 600 seconds to 3600 seconds) it stops reading directories and in the logs it just errors out with "Unable to open directory" over and over for the remaining directories.

It successfully indexed 42,094 items in 218 different directories, although that's probably only about 10% of the overall volume.

Has anyone seen this sort of behavior before? I've taken a packet capture and the traffic stops between the indexing server and the file server right after the last successful file is fully read, so it doesn't seem like it's a response that the file server is sending it that causes the indexer to error out.