I think a lot of folks are seeing this at the moment, and not just with OPNG...

As far as I can determine from a quick look at my own returned tasks, it seems there's a delay of about 24 hours between the possibility of validation and validation actually happening. There may then be a further delay before the tasks get assimilated, but that's a different issue...

Why "possibility of validation"? -- for OPN1/OPNG (which use "Adaptive Replication"), one initial task is sent out unless the host claiming it isn't yet entitled to validation without wingman, in which case a second task is queued at once; if there are two tasks sent out more or less together, validation won't occur until both have been returned. If one of them fails, a retry goes out and the possibility of validation gets pushed further into the future.

Whilst I can't tell whether that is the cause of your backlog (as - unlike most other BOINC projects - users can't [easiiy] examine other users' results to try and help with troubleshooting), I think there 's a fair chance it's likely to be at least one reason! I've just checked on my most-delayed validations, and every single one of them was an immediately issued _1 task; the matching _0 task was User Aborted after a couple of days; I now have to wait for another wingman, so my task appears to have been stuck for three days instead of one...

Cheers - Al.

P.S. I'm just a user who likes to analyse various statistics about my processed tasks, including what information might be available about wingman progress. I have no insider information...