Why is the deadline for port 11002 (GFN262144) tasks so short? The oldest pending task is just 26.5 hours old as of this post.
For example, one of my hosts was issued the candidate 142924^262144+1 on Thursday. I stopped the client for the challenge (GeneferCUDA uses too much CPU on Linux) but now I find that the candidate was allocated to another user 24 hours ago!! The prpclient log shows a normal exit so I'm sure I didn't abort it even unintentionally.
These candidates are trivially over 1M digits so the deadline should be 14 days, right? It's not such a loss for tasks intended for GPU, but I think I had some CPU tasks in progress on the Windows side of this host. Even the x64 app takes half a day to PRP test one of these so the waste of CPU time could be significant.