Guest Posted September 29, 2005 Report Share Posted September 29, 2005 It'd be great to add a feature which allows any peers which have lots of hash errors to be banned.A feature which bans peers that have over x errors so that the client is not always downloading bad data. Link to comment Share on other sites More sharing options...
Zebra Posted September 29, 2005 Report Share Posted September 29, 2005 Yeah agree with that. Could be an option whether it should be a permanent ban or for x hours. Link to comment Share on other sites More sharing options...
Guest Posted September 29, 2005 Author Report Share Posted September 29, 2005 Yup, agree too. I just hope the default criteria for banning is made intelligently. I wouldn't have a clue what numbers to plug in. Link to comment Share on other sites More sharing options...
r00ted Posted October 3, 2005 Report Share Posted October 3, 2005 I dont even care if it's a "ban" as long as Utorrent is intelligent enough to "drop" the bad hash/data, and atleast inform the user of the IP, and time the (hour:minute:second, using local time) bad data was received. Link to comment Share on other sites More sharing options...
Guest Posted October 3, 2005 Author Report Share Posted October 3, 2005 I think it definately needs to be a ban.Reason i posted this was due to one of my linux iso's having about 400 bad packets of data. Which worked out to be about 1 gig lost.Even if the ban is for an hour or 2 that'd alteast stop certain clients constantly filling the connection with bad data. Link to comment Share on other sites More sharing options...
xerya Posted October 4, 2005 Report Share Posted October 4, 2005 Not only should it be a ban, it would be nice to have a way to relate the number of allowed hash fails to the piece size, so users could decide how much of bad data they allow one peer to send before banning it.I would also very much appreciate a notification of repeated hash fails. On private trackers where ratios limits are quite strict, it isn't nice to notice that you have downloaded five times the size of torrent. Link to comment Share on other sites More sharing options...
winMX_67 Posted October 4, 2005 Report Share Posted October 4, 2005 Yeah I agree with this too. Link to comment Share on other sites More sharing options...
Drakkhen Posted October 5, 2005 Report Share Posted October 5, 2005 Definitely need this feature. Peers sending bad data need to be banned permanently. Just about the only feature I miss from my prior Java-based bittorrent client. Link to comment Share on other sites More sharing options...
vurlix Posted October 5, 2005 Report Share Posted October 5, 2005 This is implemented in 1.1.4, albeit in a simplistic way Link to comment Share on other sites More sharing options...
winMX_67 Posted October 5, 2005 Report Share Posted October 5, 2005 This is implemented in 1.1.4, albeit in a simplistic wayYESSS!!! Link to comment Share on other sites More sharing options...
Teloriun Posted October 5, 2005 Report Share Posted October 5, 2005 Good to hear it's being implemented...On private trackers it's not so much of a problem, but on publics..sheesh. Hash fail city. Link to comment Share on other sites More sharing options...
Knight Posted October 5, 2005 Report Share Posted October 5, 2005 good to hear, I lost acouple of gigs on my ratio yesterday because of bad data Link to comment Share on other sites More sharing options...
UTorrentUser Posted October 5, 2005 Report Share Posted October 5, 2005 Not only should it be a ban, it would be nice to have a way to relate the number of allowed hash fails to the piece size, so users could decide how much of bad data they allow one peer to send before banning it.I would also very much appreciate a notification of repeated hash fails. On private trackers where ratios limits are quite strict, it isn't nice to notice that you have downloaded five times the size of torrent.And just how much bad data do you think people would allow? Link to comment Share on other sites More sharing options...
Drakkhen Posted October 5, 2005 Report Share Posted October 5, 2005 This is implemented in 1.1.4, albeit in a simplistic wayDoesn't matter how, as long as it stops getting flooded with hash fails, it's all good! Link to comment Share on other sites More sharing options...
ScubaSteve Posted October 5, 2005 Report Share Posted October 5, 2005 would be a good idea. also on the sticky post at the top of your currently upcoming features would it be possible to put an ETA. Link to comment Share on other sites More sharing options...
Smoovious Posted October 7, 2005 Report Share Posted October 7, 2005 And just how much bad data do you think people would allow?It shouldn't ban on the occasional bad packet... bursts of noise and buffer overruns during transit still happen from time to time.1%, I'd find acceptable. You're still getting plenty of useful data.50%? unacceptable... ban... 20%? ... probably still ban...probably would even ban on 10%...Should have some tolerance for bad packets tho. Bad packets aren't as common as they used to be pre-HST/Telebit days, but they still happen although much more rarely.-- Smoovious Link to comment Share on other sites More sharing options...
ICleolion Posted December 10, 2005 Report Share Posted December 10, 2005 So how does this feature work then and has it been improved in versions after 1.1.4? Link to comment Share on other sites More sharing options...
atis Posted April 22, 2007 Report Share Posted April 22, 2007 Very bad implementation. It bans hosts, not caring in what time scale hash errors occur. It keeps constantly banning peers with really good speed (~400kB/s), even if i add them manually, after 10 or 15 minutes they are banned. I think, at first ban should occur at some ratio, not fixed 5 hash fails. Plus amount of allowed fails should be configurable, plus preferably the peers i add manually shouldn't ever be banned. I'm very annoyed, because also the old version without this feature isn't available. Link to comment Share on other sites More sharing options...
ICleolion Posted April 22, 2007 Report Share Posted April 22, 2007 a few good ideas spread over a couple of threads. Could do with someone spec'ing an overhaul for the hashfail/ban system, in a new thread for the devs to take a gander at. Link to comment Share on other sites More sharing options...
CodeRed Posted April 23, 2007 Report Share Posted April 23, 2007 Why is a peer sending you pieces which fail the hash? Link to comment Share on other sites More sharing options...
atis Posted April 23, 2007 Report Share Posted April 23, 2007 Actually i have no clue. My first idea was, just some network problems, however it should be fixed first at TCP level. So next guess would be bad implementation of bittorrent protocol. The peer i have problems is rtorrent (libtorrent based). However most of data i get from it are in good shape, so i would prefer failures just to be ignored (and move on to next chunk) Link to comment Share on other sites More sharing options...
Ultima Posted April 23, 2007 Report Share Posted April 23, 2007 Banning based on a ratio of good:bad pieces should be implemented for the next version. This should minimize banning of peers who have behaved well outside of the few bad pieces they sent you. Link to comment Share on other sites More sharing options...
ICleolion Posted April 24, 2007 Report Share Posted April 24, 2007 This ratio will also be configurable. So upto you how much wasted data you are willing to put up with. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.