RE: Dead Torrents

From: Ivan Kwok <ivan_at_mediadefender.com>
Date: Tue, 24 Apr 2007 11:32:14 -0700

Yes, that's exactly what I'm concern because it's already maxing out the CPU in current state.
For now, I will just disable the moving of 'dead' torrents. I'll look into it after the Miivi crisis ;)

-Ivan

-----Original Message-----
From: Ben Grodsky
Sent: Tuesday, April 24, 2007 11:07 AM
To: Tabish Hasan; Ivan Kwok; Ben Ebert; Jay Mairs
Cc: Ty Heath; qateam
Subject: Re: Dead Torrents

Adds overhead for check of dead torrents.

Ty and Ivan - what do you think?



----- Original Message -----
From: Tabish Hasan
To: Ivan Kwok; Ben Ebert; Jay Mairs
Cc: Ty Heath; Ben Grodsky; Randy Saaf; qateam
Sent: Tue Apr 24 10:02:24 2007
Subject: RE: Dead Torrents

I thought about it some more….

We can just do a similar check on our dead torrents…but reversed. We would check every set amount of time…for any seed/peer #’s. If they report #’s, then we would move them back into the protection_enabled section.

 

-TH

 

________________________________

From: Tabish Hasan
Sent: Monday, April 23, 2007 5:11 PM
To: Ivan Kwok; Ben Ebert; Jay Mairs
Cc: Ty Heath; Ben Grodsky; Randy Saaf; qateam
Subject: FW: Dead Torrents

 

Ivan,

 

Regarding our dead torrent process, it seems like we’re still removing many torrents with high seed/peer #s. On the attached sheet, you’ll notice that out of 20 randomly picked “dead” torrents…several (like 6) of them are very much alive and would download w/in 2 hours.

 

Any ideas of what can we do to decrease the false positives we remove from our system? I don’t know why it’s labeling this many as dead considering we now wait 48 hours.

 

-TH

________________________________

From: Daniel Lee
Sent: Monday, April 23, 2007 4:53 PM
To: Tabish Hasan
Cc: Jonathan Perez
Subject: Dead Torrents

 

Here’s the uTorrent raw data for the dead torrents I downloaded.

Received on Fri Sep 14 2007 - 10:55:51 BST

This archive was generated by hypermail 2.2.0 : Sun Sep 16 2007 - 22:19:45 BST