Thanks for this info - could you file a bug for it - attaching this  
image?  http://bugs.freenetproject.org/

Ian.

On 8 Apr 2006, at 16:52, Thomas Bruderer wrote:

> Hi devl
>
> was surprised how good 0.7 alpha is, wasn't that enthuastic before  
> I testet it.
>
> I also tried fuqid today.
>
> My test enviroment: have a server well connected to the darknet,  
> have some big neigbors, and have several connections. Downloads are  
> fine and quite fast. I am connected from my client machine with  
> only one connection to my server. So there are no big influences  
> from others, even if some requests  come in.
>
> I now tested an insert into freenet 0.7 and its horrible, well  
> horrible is an understatement. I have 2KB upstream in average.  
> Priority Maximum.
>
> I attached the Bandwithgraph, as you see there are peeks the peaks  
> are limited by my bandwith (30KB out) and happen to be around 15 -  
> 17 seconds apart. so this are the real inserts. they take here  
> around 1 second, then there is almost NO traffic at all, as you can  
> see. I dont know if fn05 have the same upload problem, but here it  
> is obvious that it just doesnt use the given bandwith. This makes  
> the inserts for me 15x times slower than it should be, thats not good.
>
> FN 0.7 says: Inserts 4. (maybe you try too few inserts the same time?)
>
> BTW: I tried more than one insert the same time, nothing changed, I  
> tried to lower the priority, it got worse which is ok since thats  
> what the priority should be like...
>
> I am not the only one mention this peaks, there was already a  
> discussion on IRC and a similar pic to mine is inserted into  
> darknet... Its still alpha, so plenty of time to change it, but it  
> should be really adressed, since the inserters are probably  
> important users :)
>
> cu
>
>
> <graph.png>
> _______________________________________________
> Devl mailing list
> Devl at freenetproject.org
> http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl


Reply via email to