First of all, greetings and congratulations to whoever developed Ktorrent - the client works great, the code looks elegant.

I have two questions and one remark, which likely do not belong to one thread:

***

A.

I have a linux setup, and a windows one. On the former I mean to use ktorrent - on the latter utorrent. Whenever I try to resume data created by ktorrent, with utorrent - it first seems to restart the job from where ktorrent finished, only to choke on some unseen before "Hard Disk Overload 100%". The utorrent process than becomes unresponsive (even to a taskkill /F), the hard drive is being written onto and windows won't boot but with a hard restart.

The hard drive I am using to store torrent data is a Samsung HD753LJ - the hard drive performance is hardly the issue, in my view. The filesystem is NTFS (ntfs-3g mounted).

Is there something in ktorrent's file allocation setup which troubles utorrent. Do You have any idea how I can make the two work together.

***

B.

Through the log viewer plugin I noticed a redundant 1+ TB "wasted". The line was reported over and over - what does that mean.

***

C.

There's a minor feature I miss from utorrent. Using private trackers I often discriminate among them: I sort my torrents (within the main screen), according to the tracker used. In utorrent there's such a column, it's missing in ktorrent.

In addition to the above: I installed ktorrent through ubuntu's repository (launchpad's experimental, I am guessing - KDE 4.2's). The help-about reports ktorrent version 3.2 (no release candidate mention), yet I didn't notice the plasmoid anywhere ... did I miss that.

47YYFFH wrote:In addition to the above: I installed ktorrent through ubuntu's repository (launchpad's experimental, I am guessing - KDE 4.2's). The help-about reports ktorrent version 3.2 (no release candidate mention), yet I didn't notice the plasmoid anywhere ... did I miss that.

3.2 is not released yet, sure it isn't 3.1.2 ?

47YYFFH wrote:First of all, greetings and congratulations to whoever developed Ktorrent - the client works great, the code looks elegant.

I have two questions and one remark, which likely do not belong to one thread:

***

A.

I have a linux setup, and a windows one. On the former I mean to use ktorrent - on the latter utorrent. Whenever I try to resume data created by ktorrent, with utorrent - it first seems to restart the job from where ktorrent finished, only to choke on some unseen before "Hard Disk Overload 100%". The utorrent process than becomes unresponsive (even to a taskkill /F), the hard drive is being written onto and windows won't boot but with a hard restart.

The hard drive I am using to store torrent data is a Samsung HD753LJ - the hard drive performance is hardly the issue, in my view. The filesystem is NTFS (ntfs-3g mounted).

Is there something in ktorrent's file allocation setup which troubles utorrent. Do You have any idea how I can make the two work together.

Don't know, you can disable preallocation, maybe that makes the problem go away. Anyway, this probably is more of a µTorrent problem, it should be able to handle data downloaded by another client.

Through the log viewer plugin I noticed a redundant 1+ TB "wasted". The line was reported over and over - what does that mean.

It is possible that some data is downloaded twice or even more, but it never should get to 1 TB.

There's a minor feature I miss from utorrent. Using private trackers I often discriminate among them: I sort my torrents (within the main screen), according to the tracker used. In utorrent there's such a column, it's missing in ktorrent.

We will see for 3.3

I am not sure the issue belongs with utorrent - but I trust Your opinion is more informed than mine. I didn't have preallocation on, during my first tests - I've set it on now (on ktorrent). Full preallocation on torrent creation.

As for the log viewer, I must correct myself: the data isn't reported as "wasted" but as "unnecessary data".

The log file, which comprises a few hours activity within a day, is 5.3 MB large.

I may post portions of it, without IP addresses, but this extract will suffice:

Code: Select all

Mon Feb 2 00:00:34 2009: Unnecessary piece, total unnecessary data : 16.0 KB
Mon Feb 2 01:24:06 2009: Unnecessary piece, total unnecessary data : 1.05 GB


I now realise I posted above a "1 TB wasted" - the right figure is 1 GB, sorry.

So ... I am not sure what happened, but a 4 GB torrent was set back by its whole progress in ktorrent, having been started in utorrent. I can't say I will do my utmost to reproduce the issue ( ), but I will surely keep a closer look at things.

Importing torrents (.torrent files and torrent data) I see that ktorent's status bar updates the "Transferred" "Down" number. Such data isn't obviously reported to the tracker, but it's nonetheless odd to see the number increase on importing torrents - oh well, not really an issue.

I thought I'd post a follow up on the issue.

I've pinpointed the "unnecessary data" to one torrent. Resuming it in utorrent, under windows, was problematic to say the least. Utorrent is set so that a 500MB cache in write and read mode is enabled.

Whenever data is downloaded, I can see pieces being finished, yet despite the 1MB/s download rate, the cache isn't flushed. I tried to do so, by quitting utorrent and thus freeing its used memory: the process hung, with the above mentioned 50% CPU usage. After a pause, and a request to reboot - I had to manually reset the whole box.

I also witnessed an earlier BSOD, during transfer for that particular torrent.

ktorrent - on the other hand - just reports this rather high "unnecessary data". Had I not checked its log, I would have never told there was an issue - until time proved I wasn't actually downloading anything, and progress stalled.

So: I am thinking there's something wrong with either a) my RAM or b) my hard drive. I would exclude the former, as I haven't noticed any odd behaviour using other applications. The hard drive will be scanned and a SMART report generated: I fear bad blocks may be spreading.

ktorrent's behaviour, in this circumstance, was perhaps too subtle: there was no warning there were such high amounts of "unnecessary" data downloaded.

47YYFFH wrote:ktorrent's behaviour, in this circumstance, was perhaps too subtle: there was no warning there were such high amounts of "unnecessary" data downloaded.

Unnecessary data has nothing to do with whatever µTorrent did, it happens when two or more peers get assigned to the same chunk. Requests for the same pieces can then be sent to multiple peers, if one of them then sends the piece, we send cancels to the rest, but they might already be sending this piece (and there is no way to interrupt the transfer of a piece), resulting in this unnecessary data.

It could also be caused by buggy or malicious clients.

I wasn't implying utorrent did "anything", I supposed the "unnecessary data" was a consequence of the fact that ktorrent couldn't physically allocate the pieces and kept redownloading them. You are telling me that is not the case.

Note - Private tracker (hdb*): few torrent clients allowed (azureus, utorrent, transmission, ktorrent, bittorando, rtorrent), and arguably paranoid checks against malicious clients.

My last point wasn't very clear (as is the rest of this thread, I apologise): I think the client could benefit from a "notification" when large amounts of data turn out to be "unnecessary" - 1 GB, but even 100 MB are a large amount, by my standard. It's a feature request, in short.

Sorry for the mess, thanks for the explanations.

I have the same problem. On private tracker, I´m downloading just from 1 seed and I´m the only 1 leech. On single torrent I have more than 3 GB unnecessary data.
For example

Code: Select all

"Ut jan 12 19:33:34 2010: Unnecessary piece, total unnecessary data : 3,1 GiB"

The seeder is using uTorrent 1.8.4.0, I have KTorrent 3.2.4.0. I read your posts, but I have no idea how to fix it.

Thank you for any help.