Since XoloX starts downloading from the beginning and ending of the file at the same time, there is no way to tell if the ending will match the beginning. If you get a corrupted file at the end then you will only resume from the same corrupted file, which leads to huge segments coming together and not matching up.

With crc or some other file hash meta-data this could be avoided, but currently, using only file size, this is only marginally avoidable...

Since you're downloading a movie, or ISO... look in yoru resumeinfo.txt and look for the file name that shows up most often in the HOSTS sections of that JOB and use it in the search terms section of the JOB. Using an exact, or nearly exact, name to search with will reduce your hits, but you will most likely get many fewer bad segments.