Jump to content

flak

Established Members
  • Posts

    12
  • Joined

  • Last visited

Everything posted by flak

  1. This and ...which is what I thought. And yet I am certain that RC2 (in my case) is prioritising TCP over uTP and not the other way round. For example, peers that I am connected to begin with uTP but after time revert to TCP on the same torrent, say, after a tracker update. The IP address of the peer is unchanged, as is the port. The vast majority are on UT 1.83+ as well. Stop and restart torrents and/or uTorrent and the connection to those same peers reuses uTP once again. So unless uTP is being disabled 'in situ' this shouldn't be happening, right? Is it possible to be choked by uTP connections thus forcing a switch to TCP? Perhaps that is why stopping/restarting UT eradicates this problem temporarily, even if RC1 worked fine?
  2. Upgrading from RC1 to RC2 seems to have produced a weird phenomenon for me with regards to the uTP protocol. Whereas previously with bandwidth management enabled connections seemed to favour uTP over TCP (bt.transp_disposition set to 15) and worked a treat, RC2 appears to do the opposite after a period of time. The effect for me is being connected to 80 -90% of peers on TCP with download and upload speeds reduced to a crawl. Restarting UT solves the problem, albeit temporarily. Setting net.calc_overhead to false helps a little but not much. Isn't uTP prioritised as a rule of thumb or is it just a crap shoot?
×
×
  • Create New...