1,3x: Session Transfer Vs Actual Download (Fixed in 1.40)
Re: Session Transfer Vs Actual Download
This is really interesting - and it would explain all of my problems this week. I've upgraded from a 4Mbps to 10Mbps line and each day I've been hitting traffic management limits though actual downloads are far lower than the limits. Never had these probs before. I'll try v1.22 for the next few days.
Re: Session Transfer Vs Actual Download
there's a possible fix for this as of r6865. I'd appreciate people seeing this behavior test out a nightly and report back.
-
- Posts: 4
- Joined: Tue Sep 30, 2008 6:56 am
Re: Session Transfer Vs Actual Download
Yeah the nightly has seemed to fix it.
Top effort
Top effort
Re: Session Transfer Vs Actual Download
nice work
thanks
what was the cause?
thanks
what was the cause?
Re: Session Transfer Vs Actual Download
I'm also trying the latest nightly. First observations seems to be good.
After an hour or two of downloading, I have doubts. I do not see the file growing as expected. Downloading at an average of 50kbps, the file grew by about 80MB and in the statistics window I see the session download has increased by 180MB.
Or could it be because I was using build r6868?
IN same time with 1.22, more than 200MB got downloaded.
After an hour or two of downloading, I have doubts. I do not see the file growing as expected. Downloading at an average of 50kbps, the file grew by about 80MB and in the statistics window I see the session download has increased by 180MB.
Or could it be because I was using build r6868?
IN same time with 1.22, more than 200MB got downloaded.
Re: Session Transfer Vs Actual Download
Thanks for working on this. I just downloaded build 6865 and unfortunately overnight still saw the same behavior - 800 meg session transfer and 200 meg on the torrent. I think perhaps it's less bad than before, but it's definitely not solved. After blowing my limit last week i went back to 1.22 and the session transfer always matched the torrent(s) almost exactly, so there's definitely still a regression somewhere.
-
- Posts: 1
- Joined: Wed Oct 08, 2008 10:53 pm
Re: Session Transfer Vs Actual Download
hi i've just regesterd here to report the same problem i was using 1.34 and 512k dsl connection "the internet in saudi arabia are very expensive" and i change to 1.22 now and it seems to work normally.
-
- Transmission Developer
- Posts: 3142
- Joined: Fri Jan 13, 2006 8:08 pm
Re: Session Transfer Vs Actual Download
drmalfares: as stated a few posts up, can you try a nightly build to see if this known issue is fixed? http://transmission.xpjets.com/
Re: Session Transfer Vs Actual Download
the previous nightlies help a little, but it's not a full fix. (not even a half fix, really). The good news is that we're working on this bug now and have it mostly tracked down. I'll keep you posted...
-
- Posts: 695
- Joined: Wed Oct 31, 2007 6:11 am
- Location: Australia
Re: Session Transfer Vs Actual Download
Good to see this being worked on. I've been testing the nightlies, first time I saw this problem was on r6794 and OS X 10.5.4, latest version I've tried is r6880 with OS X 10.5.5.
Re: Session Transfer Vs Actual Download
r6882 has another patch that may improve behavior. As with the previous one, user feedback here would be much appreciated.
Re: Session Transfer Vs Actual Download
At a glance, it seems to significantly improve behavior. However, there still seems to be a bit of extra data being downloaded; I've downloaded 173.3 MB and the session transfer is 188.5 MB.
Also, I noticed when testing: when Transmission is completely idle (i.e. absolutely no torrents active/downloading/uploading,) the download/upload speeds go up to max 5 kb/s for a few seconds at a time at random intervals. Is this a bug, or is it some kind of overhead (Transmission checking the status of paused torrents?)
So this bug was non-existent before 1.30 - is it not possible to just revert some of the code to how it was in 1.22?
Also, I noticed when testing: when Transmission is completely idle (i.e. absolutely no torrents active/downloading/uploading,) the download/upload speeds go up to max 5 kb/s for a few seconds at a time at random intervals. Is this a bug, or is it some kind of overhead (Transmission checking the status of paused torrents?)
So this bug was non-existent before 1.30 - is it not possible to just revert some of the code to how it was in 1.22?
Re: Session Transfer Vs Actual Download
This new snapshot seems to be losing way less data so far. I'm only 50 megs in, but you usually notice something by now. I'll post again if i notice anything serious overnight.
One thing i am noticing, though, is that the speed listed on the torrent is a lot more jumpy. The total session speeds seem relatively consistent (though even they seem to jump a bit more than 1.22) but the individual torrent speeds and its ETA go all over the place. This doesn't appear to be affecting the actual speed of the download, though (i think it is actually downloading at the global rate, not the jumpy torrent rate).
Also i don't think my turtle is working any more, or at least - not as rigidly as it used to.
Edit: actually... the turtle might be working, but because the reported speeds are so jumpy it doesn't "feel" like it is - i wonder if perhaps the back end is just updating the UI too frequently with precise speeds instead of smoothing it out.
One thing i am noticing, though, is that the speed listed on the torrent is a lot more jumpy. The total session speeds seem relatively consistent (though even they seem to jump a bit more than 1.22) but the individual torrent speeds and its ETA go all over the place. This doesn't appear to be affecting the actual speed of the download, though (i think it is actually downloading at the global rate, not the jumpy torrent rate).
Also i don't think my turtle is working any more, or at least - not as rigidly as it used to.
Edit: actually... the turtle might be working, but because the reported speeds are so jumpy it doesn't "feel" like it is - i wonder if perhaps the back end is just updating the UI too frequently with precise speeds instead of smoothing it out.
-
- Transmission Developer
- Posts: 3142
- Joined: Fri Jan 13, 2006 8:08 pm
Re: Session Transfer Vs Actual Download
Jumpy as in how much? A few kb's? Or more? If it's a few kb's, that might be the overhead, which wasn't shown before but now is.
Re: Session Transfer Vs Actual Download
There is a joke in here somewhere.alison87 wrote:Also i don't think my turtle is working any more, or at least - not as rigidly as it used to.