overlapping torrents

[dcouzin]

When the 1.8 GB torrent I was downloading became stalled for low availability, I found a similar 0.9 GB torrent with high availability. It turns out that the 0.9 GB torrent comprised an exact file from the 1.8 GB torrent. µtorrent treated this intelligently. It immediately transferred pieces from the stalled download to the startup second download. Thus the second download began 30% complete. There is some mystery about this. 50% of the common file had been downloaded with the first torrent, and yet only 60% of this went to the second torrent. Also, as the second torrent proceeds, I am surprised that µtorrent doesn't find any more pieces to send to the stalled first torrent. Why should it be downloading exactly the pieces that the first torrent didn't send to the second torrent? Is this a flaw in µtorrent? I'll know for sure when the second torrent reaches 50% completion, hopefully in about 6 hours.

Of course my other question about these "similar" torrents is what they contain. I believe the question can be posed in a general way that doesn't break the guiding rule of the forum. The 0.9 GB torrent consists of one big .tar file. The 1.8 GB torrent consists of that same 0.9 GB .tar file and also 20 equal size small files which are either .rar or .rxx, the x's being digits. I have the suspicion that the 1.8 GB torrent is redundant: that the 20 small files contain the same data as the one big file. In that case the 1.8 GB torrent is a poor one, and it's no wonder that it has poor availability. Is my suspicion reasonable? Are there such redundancies in torrents?

It is a great adventure downloading (and uploading) with µtorrent. I thank all who've worked to build this elegant place.

[Firon]

If you download the same files on two torrents, you just wasted a bunch of bandwidth. Each torrent will not interact with the other.

[dcouzin]

Firon, I guess I failed to convey what happened. I attempted to download a 1.8 GB torrent and the downloading froze at about 50%. Availability was just 50%.

So I sought another similar torrent which showed high availability. It _turned out_ that the second torrent consisted of the biggest file in the first torrent. Now with the greater availability, I'm expecting to get the file via the second torrent. The first torrent failed.

If I wasted a bunch of bandwidth it was only by trying to download the first torrent. But the first torrent download started well. The availability fraction kept growing and I was hoping it would reach 1. What should I have done in this situation?

I'm pretty sure I witnessed the interaction. In the first seconds of the download of the second torrent, I saw the "done" gauge grow from 0% to 35%. And the picture showing the acquired pieces was about 1/3 full. No download is this fast. What else could this be than µtorrent finding pieces of one torrent in another partially downloaded torrent?

[Firon]

If you continued downloading on -both- torrents at the same time, you wasted bandwidth.

Now, if you loaded those files into another torrent, it will pick up pieces because of the hash check. If you continued both, then it would not pick up pieces again (hence the bandwidth waste).

[dcouzin]

Concerning the interaction, µtorrent displays this:

size: 924 MB

downloaded: 69.2 MB

done: 35.0%

Now 69.2 MB is nowhere near 35% of 924 MB, so 254.2 MB of this torrent was borrowed from the earlier downloaded torrent.

My question was: why just 254.2 MB, since there was 486 MB of that same file already downloaded in the first torrent?

Is there a way to manually move the 486 MB of the partially downloaded first torrent into the download of the second torrent, to speed it up (and save bandwidth)?

Thanks.

[Firon]

Incomplete pieces cannot be used. The data might've also been bad. Who knows.