Opened 9 years ago

Closed 9 years ago

Last modified 9 years ago

#4975 closed Bug (wontfix)

verification of large torrents needs to checkpoint progress

Reported by: jpritikin Owned by:
Priority: Normal Milestone: None Set
Component: Transmission Version: 2.60
Severity: Normal Keywords:
Cc:

Description

I'm trying to serve a giant torrent (800GiB and growing) using an Asus RT-N16 router:

http://www.dd-wrt.com/wiki/index.php/Asus_RT-N16

This router only has USB 2.0 so disk throughput is rather slow. The initial verify of the torrent takes about 24 hours. The problem is that, if anything goes wrong (e.g. a brief power outage), transmission wants to start over from 0% verified.

It would be most helpful if transmission remembered how much of the torrent was verified during the verification process.

Change History (3)

comment:1 Changed 9 years ago by jordan

  • Resolution set to wontfix
  • Status changed from new to closed

IMO the point of verification is to throw out any assumptions and start from scratch. What if something happened to the file, or to the disk, after the last verify pass?

I understand the use case you're describing but think the solution you're suggesting would solve one problem at the cost of creating a worse one for other users.

comment:2 Changed 9 years ago by cfpp2p

What we might need is a proper fix of #1465 that would precisely verify file by file therefore solving this ticket (#4975) as each file would be verified as proceeding. #4262 will also be immediately simplified as well. Patch for one --> fix for three would work for me.

comment:3 Changed 9 years ago by jpritikin

What if something happened to the file, or to the disk, after the last verify pass?

By that logic, why doesn't transmission re-verify already verified torrents after every restart? In other words, why is full verification sticky but partial verification is not sticky?

Note: See TracTickets for help on using tickets.