![]() ![]() I was watching the debug log while this was happening, and while I do see a lot of the "do_purge flag because we got an ERANGE, EMSGSIZE, or ENOTCONN error", they were happening regardless of the throughput of the torrent - I mean, someone should still look into this, but it's not related to the performance problems. Unapuse the public torrent: same story.Unpause the private torrent: it finished the remaining 80% of the torrent always fully utilizing bandwidth.Quit transmission, unload and load the networking module, reconnect.All the peers were found rather quick, but I was still seeing the lower speeds. Paused the private torrent, and added the public one.After that, speeds dropped to 6~10MB/s and kept there for a while. Started the private torrent, and for the first 60 seconds I was seeing the expected speeds (25MB/s, but with some variation).I gathered two large torrents, one public (4GB) and one private (6GB), both of them are well seeded (I was seeing 100 seeds for each).IDK what's causing it and it could be nothing, but best to make sure."īut (unless Google fails me) it seems like no further investigation of those errors has occurred? I see mentioned alongside performance complaints in downloads stuck at 99% when torrent piece size isn't cleanly divisible by 2^14 #3111 and macOS: 'Libevent got an error!' #4046 - where said "That one's probably worth opening a separate ticket for.does not appear in the Transmission 3 logs at all.What stands out is that the Transmission 4 log is absolutely littered with this error:ĭo_purge flag because we got an ERANGE, EMSGSIZE, or ENOTCONN error I spent some time looking at the debug-level log for the same ~1k private torrents, switching between Transmission 3 and Transmission 4 - and still observing significantly higher transfer rates on 3 and the sort of on-again, off-again speeds on 4. Possible avenue to look into?Įnabling PEX -> Same issue after 10 minutes, only 1 peer flickering in and out at 130 KB/sĮnabling PEX+DHT -> Speeds back to normal an 60 peers found with close to 60 connected 10MB/sĮnabling DHT only -> Speeds back to normal an 60 peers found with close to 60 connected 10MB/sĭisabling them again, restart and start the torrent again -> Speeds at 10 MB/s as my IP:port is still in their tables.įWIW, I exclusively use Transmission with private torrents, so I have little experience with normal performance on public torrents or perspective to offer about PEX, DHT, etc. PEX/DHT/LPD off -> I am only getting one peer, and bursts of activity on and off at 130 KB/s. As a last thing from my end, I tried downloading the Ubuntu torrent with PEX/DHT/LPD off. It did not impact my speeds, however.Įdit 3: Scrap this edit, it's late and not explaining myself correctly! Try the above with the same settings as me, and let's see how it performs. ![]() Although for some reason UPnP fails although it is enabled on my Orbi router. This is the full debug log for the download of the Ubuntu torrent, my speeds were consistent, at 10MB/s which is utilising the full bandwidth of my connection. ![]() Note: You may need to scrub your home directory name in the logs Once the torrent completes, press save at the bottom left and upload here.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |