[freenet-support] wasting bandwidth?

2002-11-13 Thread Shawn Yarbrough
My node uses about 30% of available CPU and less than 10% of available
bandwidth.  Here are my freenet.conf bandwidth limits:

bandwidthLimit=1048576 (this line should be ignored, right?)
inputBandwidthLimit=1048576
outputBandwidthLimit=1048576

This is for a 10BaseT full-duplex Ethernet connection to the internet.  With
these settings I only see about 500Kbit/s to 2Mbit/s bandwidth usage.  In
fact the only way I got bandwidth above 500Kbit/s was by messing with
maximumThreads.  Unfortunantly raising maximumThreads to 200 or above makes
my system unable to fork so I had to reboot.  Now it is set at 150.  There
were lots of thread creation errors in freenet.log afterwards.

Before I messed with maximumThreads my number of open connections was almost
always pegged at 48.  Now it is almost always pegged at 60.  I feel this
system will easily handle 200 or more open connections if somebody can help
me configure it right.

There is plenty of RAM in this system, 768 MB.  I can provide Freenet
diagnostics and logs if needed.  This is Freenet build 535.

How can I really crank up this underused node?


Shawn Yarbrough
[EMAIL PROTECTED]


___
support mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support



[freenet-support] Full albüm mp3 rjqs

2002-11-13 Thread Girl Amlani
Mp3sa yine bir ilki gerçekleþtiriyor: Klip arþivi!
Full albüm ve single parçalar mp3 halinde!
Arayýpta bulamadýðýnýz bütün parçalar için birde sitemize bakýn: http://www.mp3sa.com

Full Turkçe Album 
Full Yabancý Album 
A-Z Yerli Mp3 
A-Z Yabancý Mp3 
En Iyý 20 
Yerli Výdeo Klýp 
Yabancý Výdeo Klýp 
Yerli ve Yab. Arsýv

Hepsine birden ulaþabileceðiz tek bir adres var
http://www.mp3sa.com



.¦š+¶f¢–)à–+-²êi¢»_­ç§zÚk¢7œ¶Šà†ÛiÿøZÂGëyéÞ¶šèç-¢¸?r›Šæj)fjåŠËbú?²êi¢»


Re: [freenet-support] wasting bandwidth?

2002-11-13 Thread Matthew Toseland
On Wed, Nov 13, 2002 at 09:22:00AM -0600, Shawn Yarbrough wrote:
 My node uses about 30% of available CPU and less than 10% of available
 bandwidth.  Here are my freenet.conf bandwidth limits:
 
 bandwidthLimit=1048576 (this line should be ignored, right?)
 inputBandwidthLimit=1048576
 outputBandwidthLimit=1048576
 
 This is for a 10BaseT full-duplex Ethernet connection to the internet.  With
 these settings I only see about 500Kbit/s to 2Mbit/s bandwidth usage.  In
 fact the only way I got bandwidth above 500Kbit/s was by messing with
 maximumThreads.  Unfortunantly raising maximumThreads to 200 or above makes
 my system unable to fork so I had to reboot.  Now it is set at 150.  There
 were lots of thread creation errors in freenet.log afterwards.
 
 Before I messed with maximumThreads my number of open connections was almost
 always pegged at 48.  Now it is almost always pegged at 60.  I feel this
 system will easily handle 200 or more open connections if somebody can help
 me configure it right.
Yeah, the max # of open connections used is maximumThreads * 0.4.
 
 There is plenty of RAM in this system, 768 MB.  I can provide Freenet
 diagnostics and logs if needed.  This is Freenet build 535.
 
 How can I really crank up this underused node?
Apart from increasing maximumThreads... I don't know. You could try
setting overloadLow = 1, overloadHigh = 1, to disable the overload
handling code...
 
 
 Shawn Yarbrough
 [EMAIL PROTECTED]
 
 
 ___
 support mailing list
 [EMAIL PROTECTED]
 http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/11/02.
http://freenetproject.org/



msg02101/pgp0.pgp
Description: PGP signature


Re: [freenet-support] Whining again about bandwidth

2002-11-13 Thread Matthew Toseland
On Mon, Nov 11, 2002 at 03:27:54PM -0500, Doug Bostrom wrote:
 Greetings,
 
 534 seems to be running just great. For me there's just one remaining problem, which 
is that I still don't seem 
 to be able to get bandwidth under control. 
 
 I have things pinched off to 10K up and down, on the premise that since we're on 
ADSL here I can't allow more 
That's the combined limit or the individual limits?
 relayed data  inward than my upstream connection can push out. There's some 
confusion about whether we're 128k 
No reason not to have asymmetrical limits.
 or 384k upstream. Our service agreement says 384k, but maybe that's on a good day 
with a tailwind; most of the 
 time the upstream connection performs like 128k.  So I'm erring on the safe side and 
setting things at 10K, 
 which seems only marginally useful but ought to work and leave some overhead for 
other services to negotiate 
 connections. 
I use 20k down, 5k up, on a 512/128 cable modem (I have three nodes in
this configuration on the same machine, but they're not heavily loaded). Are
you sure you aren't running any other bandwidth hogging apps? If you set
it to 5k up, does the situation improve?
 
 Our pipe is still saturated after a few hours of freenet uptime.
 
 My spouse and I both  work at home at least 50% of the time. Yesterday I 
absentmindedly referred to shutting 
 down the freenet node here when Ann was desperately trying to get some work done. 
Big mistake! Busted! Now of 
 course when things are slow I'm getting the question you don't have that THING 
turned on again, do you?, heh.
 
 Does bandwidth control take into account relayed data, ie data transiting my node as 
a result of indirection? 
 And do you have any further tips I might apply to getting this to work? Very 
frustrating to see things working 
 so well that I can't support a persistent node anymore! 
It takes into account everything, except for traffic to non-internet
addresses.
 
 Thanks for any help that may be forthcoming.
 
  
 
 
 
 --
 Democracies die behind closed doors.
 - Judge Damon Keith 
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02103/pgp0.pgp
Description: PGP signature


Re: [freenet-support] heisenbug

2002-11-13 Thread Matthew Toseland
On Sun, Nov 10, 2002 at 11:03:54AM -0600, Shawn Yarbrough wrote:
 Two things:
 
 
 1) last night my dedicated Freenet node (which seems to still be running
 fine) printed to the shell that spawned it:
 
 PANIC! Just recreated heisenbug! ID: 7a0e982c7a0e982c

As long as you don't get more of these, this is probably not a problem.
They will occasionally happen by chance. More of them usually indicates
a JVM problem.

 
 
 2) why didn't this message go to the freenet.log ?  it's annoying that my
 console keeps getting spammed with freenet messages

It does now.

 
 
 
 I'm running build 534.  I can provide node diagnostics if necessary.
 
 
 Shawn Yarbrough
 [EMAIL PROTECTED]
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02104/pgp0.pgp
Description: PGP signature


Re: [freenet-support] Keep up the good work, but ...

2002-11-13 Thread Matthew Toseland
On Sun, Nov 10, 2002 at 11:03:36AM -0600, Robert Carroll wrote:
 Great job, freenet is getting much better.  However, fproxy access for non localhost 
addresses is still slow if you are running a busy permanent node.
Non-localhost addresses? LAN addresses, you mean, e.g.
10.x.x.x
192.168.x.x
etc?
Or do you mean you are trying to access the node over the internet? In
this case, I recommend you use an SSL wrapper program, you get
encryption and the node sees local connections. Oh, and set
publicNode=true in your freenet.conf, if other people can access it,
this disables some potentially damaging pages.
-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02105/pgp0.pgp
Description: PGP signature


Re: [freenet-support] not all startup error reported by bunnyapp

2002-11-13 Thread Matthew Toseland
On Sun, Nov 10, 2002 at 12:59:34PM +0100, [EMAIL PROTECTED] wrote:
 hi there. i upgraded freenet's .jar files to the most recent snapshot, like i've 
done a dozen times before.
 but this time i stumbled into a error after i started bunnyapp. the trayicon shows 
up, then blinks after a while, signalling that freenet could not be started. 
 when i 
 loked into the logfiles, i could not find a error report, so i launched it from the 
console. there it does not launch as well, but this time i got a error respone i 
 could 
 deal with:
 
 C:\Programme\Freenet 0.5 java -jar freenet.jar
 Routing table corrupt! Remove 'rt' files and restart.
This is now logged to the logfile. At least in the development branch.
 
 perhaps it would be wise to add this message into the logfile, too, so a user of 
bunnyapp could see this, too, without knowing how to start freenet via 
 console?
 
 after sacrifying my routing table (*sob*) freenet started up without problems - of 
course! ;)
 
 congratulations again to freenet - the propagation of keys seems to be very good, 
after inserting a file with htl=10 into my local node, i was able to retrieve 
 the same file via *several* foreign ssh public node within minutes - without even 
one dnf or rnf! (i think requesting htl was 15 each)
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02106/pgp0.pgp
Description: PGP signature


Re: [freenet-support] help please

2002-11-13 Thread Matthew Toseland
On Sat, Nov 09, 2002 at 07:33:52PM +0100, j. wrote:
 Hello
 i have the freenet internet web interface running
 But what next?
 if i click on request freesite by URI
 i get meggage as below:
 also GPL does not work
 What i do wrong?
 help files are not complete
 help please
 mail to
 [EMAIL PROTECTED]
Did you ever get this working? The correct fix is to get new seednodes,
there should be a menu item to do this on the rabbit icon?
 
 thank you
 
 Network Error
 Couldn't retrieve key: SSK@rBjVda8pC-Kq04jUurIAb8IzAGcPAgM/TFE// 
 Hops To Live: 15
 Error: Route not Found 
 
 
   Attempts were made to contact 1 nodes. 
 a.. 1 were totally unreachable. 
 b.. 0 restarted. 
 c.. 0 cleanly rejected. 
   The request couldn't even make it off of your node. Try again, perhaps with the 
GPL to help your node learn about others. The publicly available seed nodes have been 
very busy lately. If possible try to get a friend to give you a reference to their 
node instead. 
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02108/pgp0.pgp
Description: PGP signature


Re: [freenet-support] reference suicide

2002-11-13 Thread Matthew Toseland
On Mon, Nov 11, 2002 at 08:27:17AM -, Vitenka - Zen wrote:
  Well, my rt files haven't exploded this time - but the databugs seem to be slowly 
munching through 
  my list of known nodes.
  Are there really only 16 nodes?
  And do we have to remove them (nodes we discard for whatever reason) from the 
list - shouldn't we 
  just put them to one side so that we can get them back?
  
 What build number?
 
 534
 
 Gah - now it dd blow up the rt nodes again and fail to boot.
There was a bug, it is fixed in the latest stable snapshots.
 
 Deleted them, working.
 
 What ARE those files?  And if deleting them is the only way to get the node back up, 
can't it do that on startup?
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02109/pgp0.pgp
Description: PGP signature


[freenet-support] How to get Kaffe to work every time

2002-11-13 Thread Rob
rm ls*, then rm rt* in the freenet directory.  Then start Freenet.


-- 
-
The Numeric Python EM Project

www.pythonemproject.com

___
support mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support



Re: [freenet-support] 528 vs. 604

2002-11-13 Thread Matthew Toseland
On Thu, Oct 31, 2002 at 02:28:45PM +0100, [EMAIL PROTECTED] wrote:
 Builds 528 and 604 have this mod, are in CVS and snapshots. Will get
 0.5.0.4 out ASAP Thursday with this change and maybe a few others.
 
 why are there still 2 different freds... what's the difference between them? new 
users of freenet hopefully have accepted that freenet is still 0.x, and thus they 
They are the stable branch and the development branch.
 had to upgrade regulary in the last few days, so they should get used to it why 
not make 6xx the current release?
 
 btw.. where can i find the latest developers' edition? are they the snapshots, or 
are they in cvs only?
From CVS, or from freenet-unstable-latest.jar in the snapshots.
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02111/pgp0.pgp
Description: PGP signature


Re: [freenet-support] Nothing but DNFs

2002-11-13 Thread Matthew Toseland
On Fri, Nov 01, 2002 at 08:43:35AM -, Vitenka - Zen wrote:
 No reply at htl 15.
 Trying htl 40  Not found.
 
 But no one can see me either.
 
 Is it posisble that the network has just gotten so 
 big (and full of holes) that an insert of 5 and a 
 request of 40 simply don't reach far enough across 
 the network?
Yes. Use HTL of at least 15 to insert. HTLs are limited to 25 currently,
so above that is pointless.
 
 There's definately a lack of flow - most of my 
 requests are timing out with '25 hosts queried - 19 
 didn't connect, 6 cleanly failed'
 
 But people seem to be using my node, and I can get 
 a number of freesites.
 
 Still getting awful performance especially if I 
 fire up frost.
 
 I also have a complaint to whoever said that 100 
 idling threads is harmless.
 Not on windows it isn't, certainly not in java on 
 windows.  Threads are rpocesses and have horrible 
 overhead.
Whatever. We can't get rid of them without huge effort. It might be
worth it someday (nonblocking I/O).
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02112/pgp0.pgp
Description: PGP signature


Re: [freenet-support] Connections running amok

2002-11-13 Thread Matthew Toseland
On Tue, Oct 29, 2002 at 04:58:00PM -0400, Doug Bostrom wrote:
 On Tuesday 29 October 2002 01:55 pm, you wrote:
 
 
 
  Use the bandwidth limiter to prevent it swamping your connection. It is
  expected to use lots (up to 120) threads - they should normally be
  almost all idling. The current situation is a product of the network
  being slashdotted, and overloaded...
 
 
 Let me emphasize once again that I've had bandwidth as low as 8k, using both 
 the combined option and the separate inbound/outbound options. Makes no 
 difference here; the pipe gets saturated. Connection is ADSL 1700/384k. Can't 
 even get a ping out when the node is up.
Hmm, that's bad. Does it still happen with 535?
 
 What does MaxNodeConnections control? It appears to have no relationship to 
 the number of inbound and outbound connections. 
 
 All re 525.
 

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02116/pgp0.pgp
Description: PGP signature


Re: [freenet-support] Keep up the good work, but ...

2002-11-13 Thread Robert Carroll
I tried adding publicNode and I couldn't see any difference.  The node is
running under windows XP.  Both computers have valid internet addresses.
Both computers are on the same subnet.  The node is configured to allow all
IP addresses mainport access, however my firewall only permits the two
computers to access fproxy or fcp.  However, with the latest release it is
alot better, but you still sometimes you have to wait up to a minute for
fproxy to respond.  Localhost always seems fast, so why not have an option
that allows all FCP and FProxy access to be treated like localhost.  Right
now, FCP and FProxy access from non-local host addresses ARE treated
differently.  Just give us the option to turn that behaviour off.  I've
asked this a couple times before and I was told to use software like squid,
putty, etc.  But I want to use mozilla and setting up squid on win xp is non
trivial, not to mention my poor server is overloaded right now.

- Original Message -
From: Matthew Toseland [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Wednesday, November 13, 2002 3:49 PM
Subject: Re: [freenet-support] Keep up the good work, but ...

 Great job, freenet is getting much better.  However, fproxy access for non
localhost addresses is still slow if you are running a busy permanent node.
Non-localhost addresses? LAN addresses, you mean, e.g.
10.x.x.x
192.168.x.x
etc?
Or do you mean you are trying to access the node over the internet? In
this case, I recommend you use an SSL wrapper program, you get
encryption and the node sees local connections. Oh, and set
publicNode=true in your freenet.conf, if other people can access it,
this disables some potentially damaging pages.


___
support mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support



Re: [freenet-support] wasting bandwidth?

2002-11-13 Thread Edgar Friendly
Shawn Yarbrough [EMAIL PROTECTED] writes:

 My node uses about 30% of available CPU and less than 10% of available
 bandwidth.  Here are my freenet.conf bandwidth limits:
 
 bandwidthLimit=1048576 (this line should be ignored, right?)
 inputBandwidthLimit=1048576
 outputBandwidthLimit=1048576
 
If you're not coming anywhere near these values, increasing them won't
change anything.

 This is for a 10BaseT full-duplex Ethernet connection to the internet.  With
 these settings I only see about 500Kbit/s to 2Mbit/s bandwidth usage.  In
 fact the only way I got bandwidth above 500Kbit/s was by messing with
 maximumThreads.  Unfortunantly raising maximumThreads to 200 or above makes
 my system unable to fork so I had to reboot.  Now it is set at 150.  There
 were lots of thread creation errors in freenet.log afterwards.
 
This is where the problem is.  The solution to get fred to work harder
for you is to just increase the maximumThreads.  The problem is that
your JVM is running into OS-defined limits on the number of
threads/processes.  Try fiddling with the ulimit command to increase
the number of processes that the user running fred is allowed.  This
should take care of the fork problems.

 Before I messed with maximumThreads my number of open connections was almost
 always pegged at 48.  Now it is almost always pegged at 60.  I feel this
 system will easily handle 200 or more open connections if somebody can help
 me configure it right.
 
It's just an OS issue limiting the # of threads.  When we get async
IO, we won't need anywhere near as many threads, and this kind of
tweaking won't be necessary.

 There is plenty of RAM in this system, 768 MB.  I can provide Freenet
 diagnostics and logs if needed.  This is Freenet build 535.
 
 How can I really crank up this underused node?
 
 
 Shawn Yarbrough
 [EMAIL PROTECTED]
 

I hope this helps,
Thelema
-- 
E-mail: [EMAIL PROTECTED] Raabu and Piisu
GPG 1024D/36352AAB fpr:756D F615 B4F3 BFFC 02C7  84B7 D8D7 6ECE 3635 2AAB

___
support mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support



Re: [freenet-support] Re: [freenet-dev] 0.5.0.1 half-released

2002-11-13 Thread Matthew Toseland
On Tue, Oct 29, 2002 at 01:28:18PM +0100, Marco A. Calamari wrote:
 At 07.01 29/10/02 -0500, you wrote:
 
  Currently rel-0-5-1 has it
 [...]
 
 A word of warning to people who may try this: port 8889 is GONE.  The
 node status stuff has been merged into port .
 
 What about distribution port ?
 
 Better, what about distribution service ???
Distribution port and service are now on by default in the development
branch, with some important bugfixes, and linked from the mainport
homepage. This will be merged across shortly (for 0.5.1).
 
 Ciao.   Marco
 
 
 -- 
 + il  Progetto Freenet - segui il coniglio bianco+
 * the Freenet  Project - follow the  white rabbit*
 *   Marco A. Calamari[EMAIL PROTECTED] www.marcoc.it*
 * PGP RSA: ED84 3839 6C4D 3FFE 389F 209E 3128 5698   *
 + DSS/DH:  8F3E 5BAE 906F B416 9242 1C10 8661 24A9 BFCE 822B +
  

-- 
Matthew Toseland
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Freenet/Coldstore open source hacker.
Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03
http://freenetproject.org/



msg02119/pgp0.pgp
Description: PGP signature


Re: [freenet-support] wasting bandwidth?

2002-11-13 Thread Greg Wooledge
Shawn Yarbrough ([EMAIL PROTECTED]) wrote:

 Unfortunantly raising maximumThreads to 200 or above makes
 my system unable to fork so I had to reboot.  Now it is set at 150.

You may need to recompile your kernel to allow more processes.
Unfortunately you didn't say what operating system you're using
(or what Java VM).

-- 
Greg Wooledge  |   Truth belongs to everybody.
[EMAIL PROTECTED]  |- The Red Hot Chili Peppers
http://wooledge.org/~greg/ |



msg02120/pgp0.pgp
Description: PGP signature


Re: [freenet-support] Whining again about bandwidth

2002-11-13 Thread Doug Bostrom
11/13/02 4:42:36 PM, Matthew Toseland [EMAIL PROTECTED] wrote:
 I have things pinched off to 10K up and down, on the premise that since we're on 
ADSL here I can't 
allow more
That's the combined limit or the individual limits?

Individual.

I use 20k down, 5k up, on a 512/128 cable modem (I have three nodes in
this configuration on the same machine, but they're not heavily loaded). Are
you sure you aren't running any other bandwidth hogging apps? If you set
it to 5k up, does the situation improve?


No other apps taking bandwidth.

Is it correct to conclude that if I'm on ADSL and I fail to take account of the 
asymmetric nature of 
the connection that problems can result? To wit:

As I understand it, some requests arriving at my node are forwarded to other nodes, 
with the results 
being passed back through my node. If my inbound connecton is 2X, and my outbound 
connection is 1X, 
this means that data going _across_ my node can arrive at my node at twice the rate it 
can leave my 
node. In other words, data can transit my site based on Freenet's method of protection 
by indirection, 
and if the results of data requests that transit my site can arrive at my site much 
faster than they 
can then leave on their way to their ultimate destination, it seems that I must base 
my inbound 
bandwidth settings strictly on my outbound speed.

Put yet another way, the trouble with the transit thing and ADSL is that since 
protocol packets 
requesting data are presumably smaller than the resulting data packets coming back, it 
seems easily 
possible that a node will happily accept and forward request packets and then attempt 
to relay far more 
resulting data packets than it's capable of dealing with if the inbound bandwidth 
settings are not 
arranged strictly on outbound connection speed.

Is this a correct interpretation, or is the bandwidth control smart enough to account 
for ADSL 
peculiarities?

I suppose another upshot of ADSL is that my local datastore can be posted to many 
times faster than it 
can be retrieved from, but that does not seem such a large potential problem as the 
transit issue.






___
support mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support