Stuck debugging

BUG FIX

Description of Bug

The logs of the renter were showing the repair loop getting stuck in the same directory. The same number of chunks would be added to the heap each time. This was an issue for files that were no longer on disk due to an inconsistency of handling the RemoteRepairDownloadThreshold in managedFetchLogicalChunkData.

Fix / Changes to Work Flow

This MR impletes 3 fixes.

1. Since managedCleanUpUploadChunk only updates the chunk stuck status when the chunk is complete, any returns in threadedFetchAndRepairChunk due to errors would result in the chunk not being marked as stuck. This MR adds in marking the chunk as stuck if there is an error with the file.
2. managedFetchLogicalChunkData now compares the chunk health to the RemoteRepairDownloadThreshold in the same way as the rest of the repair code.
3. buildUnfinishedChunks now compares chunk health against RemoteRepairDownloadThreshold when deciding whether or not to add the chunk to the heap.