0.70 / 0.71 caused the connections between peers to stop?

PUSH · September 6, 2006, 5:17am

Hello there:

0.70(and above) caused the connections between peers to stop. not lose connection, just stop.

Every peer seems to download and share the same blocks of the files, that caused many 99.7% (for example) peers and only few 100% owner.

And, the 100% owner stopped to upload anything, while still connected and still trying to upload. (no matter the 100% owner are using 0.7x or 0.6x)

That’s what we’re suffering. Thank you.

p.s. After I downgrade to 0.6x, I did get more blocks of the file, but still incomplete , so I guess it’s due to there’re still lots of 0.7x peers.

:huh:

like this..

/post-6590-1157555427.png" rel=“”>

PUSH · September 8, 2006, 5:34pm

And this, it’s another downloading job differs to above

note this:

**1) There’s a seed at least, and is still alive ( the status shows “D_d5_0” ) **

2) No upload/download from other peers, as the pink area shows (while my other job works fine).

3) Even if it’s due to the seed uploads very slow, at least there should be transmition between others, but no…

4) The 91.8% peer seems not uploading/downloading anything, that’s why I think the whole connection between these peers are dead (or almost dead).

Why??

Any opinion?/post-6590-1157736865.png" rel=“”>

PUSH · September 9, 2006, 3:46pm

And this, it’s another downloading job differs to above

After 20 hours, just 0.1%~0.2% progress

see this

/post-6590-1157816489.png" rel=“”>

Really poor condition…

(The BitLord peer got 0.3%, even better then BitComet…)

Why?

Dark_Shroud · September 10, 2006, 3:35am

0.71 tries to make greater use UDP but had some bugs in the process. Once the bugs are fixed and people upgrade the speeds will be good.

BitLord is BitComet 0.58 with a slightly differnet interface and spyware.

PUSH · September 10, 2006, 5:55am

0.71 tries to make greater use UDP but had some bugs in the process. Once the bugs are fixed and people upgrade the speeds will be good.

BitLord is BitComet 0.58 with a slightly differnet interface and spyware.

Thanks a lot.

So what would be help?

I’ve try to tell peers (via chat function) to use lower version of BitComet in order to finish this job,

But due to the chat was “partially disabled”(?) in 0.7x,

it seems no peer got my message.

And even I downgrade to 0.60 (same version as seed),

still I can’t get reasonable download speed.

seems 0.7x produced some interference between peers?

PUSH · September 10, 2006, 10:15am

Say, when 0.7x peer appears,

1)If there’re only 1 seed, it’s hardly for each peer to recieve datas from the seed directly, so they just exchange what they already have, untill nothing to offer. That makes the situation like example #1 (many 99.7 peers while the seed is still there).

2)And only if there’re more then 1 seed (and lucky enough), then this download is more easily to complete.

Although this is normal with BT, but since 0.7x appear, it became more significant, especially when the seed is not using 0.7x.

This maybe related with BC 0.71 wont accept peers and “The program thinks it is done and starts uploading”.

Any idea?

joester · September 10, 2006, 2:55pm

Say, when 0.7x peer appears,

1)If there’re only 1 seed, it’s hardly for each peer to recieve datas from the seed directly, so they just exchange what they already have, untill nothing to offer. That makes the situation like example #1 (many 99.7 peers while the seed is still there).

2)And only if there’re more then 1 seed (and lucky enough), then this download is more easily to complete.

Although this is normal with BT, but since 0.7x appear, it became more significant, especially when the seed is not using 0.7x.

This maybe related with BC 0.71 wont accept peers and “The program thinks it is done and starts uploading”.

Any idea?

Something I noticed…

In both .70 and .71, when a file reaches 100%, all uploading stops unless the task specific settings in the file properties is changed from [NO LIMIT]. I let BC run 10 hours, and nothing uploaded. I changed task specific settings, and uploading started immediately. It was a bit less than my settings, but now I can at least hope to get my U/D ratio even.

PUSH · September 10, 2006, 6:06pm

Something I noticed…

In both .70 and .71, when a file reaches 100%, all uploading stops unless the task specific settings in the file properties is changed from [NO LIMIT]. I let BC run 10 hours, and nothing uploaded. I changed task specific settings, and uploading started immediately. It was a bit less than my settings, but now I can at least hope to get my U/D ratio even.

Yes, I experence that too, but sometimes maybe it’s because the priority being low.

What’s your priority?

PUSH · September 10, 2006, 6:25pm

Progress report~

/post-6590-1158772682.png" rel=“”>

(Compare with my post#2&3)

======================================

Time passed: 20+hours 25+hours

BitComet: 90.7% → 90.9% → 91.5%

BitLord: 91.8% → 92.1% → 92.4%

BitComet: 0.2% 0.6%

BitLord: 0.3% 0.3%

======================================

so during these time, BitComet wins

But note this: peers are not grow up with the BitLord peer,

So,The BitLord (BitComet 0.58 core) peer never upload?

And the seed (BitComet 0.6) still uploads very few.

Does that means 0.7x communicates bad with lower version (when there is only 1 seed)?

PUSH · September 20, 2006, 5:23pm

To whom may concern:

since Sep 9 2006, 01:34 AM till now (Sep 21 2006, 00:33)

The progress grows from 90.7% to 93.9% (total size 4.10GB)

It took almost 12 days (287 hours) but just 3.2% downloaded!

So, as Dark_Shroud said here,

( UDP connection bug in .70/.71 )

maybe “not really a bug more of a settings issue and a problem with your router”,

but what if the amount is too large to affect from peers to ISPs,

Is it possible to reduce the speed?

Or cause the connections between peers slow or stop?

Since the problem in 0.71 / 0.72 was not solved, maybe the stuff should try think from this point of view.

Also, I can’t image why a seeder prefer to upload with such a extreme slow speed.

maybe there’s something wrong with 0.70 was not noticed yet?

someonehoi · September 22, 2006, 5:46pm

To whom may concern:

since Sep 9 2006, 01:34 AM till now (Sep 21 2006, 00:33)

The progress grows from 90.7% to 93.9% (total size 4.10GB)

It took almost 12 days (287 hours) but just 3.2% downloaded!

So, as Dark_Shroud said here,

( UDP connection bug in .70/.71 )

maybe “not really a bug more of a settings issue and a problem with your router”,

but what if the amount is too large to affect from peers to ISPs,

Is it possible to reduce the speed?

Or cause the connections between peers slow or stop?

Since the problem in 0.71 / 0.72 was not solved, maybe the stuff should try think from this point of view.

Also, I can’t image why a seeder prefer to upload with such a extreme slow speed.

maybe there’s something wrong with 0.70 was not noticed yet?

After i read your screenshot, i dont think it is a bug.

Because this situation is very normal when a file is not very popular .

That means the file is only 1 seed with very few peers.

Therefore, some seeders like to control the upload for a few day,

so that more peers which have already completed 9x% help to share the file.

If seeder dont keep the 9x% peers, the peers may gone very soon after 100%.

An other situation is that the seed file itself has error,

it cannot be uploaded to 100% no matter you wait how many days.

In my experience, BT download speed greatly depends on how many peers you have connected.

Imagine if each peer shares 1KB to you,

10 peers → you have 10KB download

100 peers → you have 100KB download

the more peers ,the faster you download.

PUSH · October 4, 2006, 4:59pm

After i read your screenshot, i dont think it is a bug.

Because this situation is very normal when a file is not very popular .

That means the file is only 1 seed with very few peers.

Therefore, some seeders like to control the upload for a few day,

so that more peers which have already completed 9x% help to share the file.

If seeder dont keep the 9x% peers, the peers may gone very soon after 100%.

An other situation is that the seed file itself has error,

it cannot be uploaded to 100% no matter you wait how many days.

In my experience, BT download speed greatly depends on how many peers you have connected.

Imagine if each peer shares 1KB to you,

10 peers → you have 10KB download

100 peers → you have 100KB download

the more peers ,the faster you download.

Thanks.

I’ve been using BitComet from 0.5x till now, downloaded more than 700+GB.

I can tell what’s normal and what’s not.

All you said is BT’s Law, that’s right,

But I still think there’re something wrong.

Because same situation happened - Peers are hard to get data from the seeder,

this time, I am the seeder.

It’s a 52MB job with around 60 peers (still increasing),

every peer (including me) stopped at 99.7% while there’s at least 1 seeder,

after 3 days, I reached 100%, so I try to seeding, and set it to upload at >5 kB/s speed.

24 hours passed, peers are still at 99.7%,

almost nothing uploaded.

meanwhile, my other downloading jobs worked ok,

Isn’t this unnormal?

By the way, since last time I’ve post,

The 4.10GB job is 95.9% now… (still 170MB left to be download)

<< << <_<