re: [freenet-support] node on a server, and at home?
Hi I'm also running a node on my private serverfarm and accessing it from home. I succesfully created ssl tunnel with stunnel from my Linux router to my Linux server for surfing freenet with my browser. My problem is that I'm having trouble to create working stunnel between Fuqid (running on WinXP) and my nodes fcp-port, Fuqid tells me that my node is overloaded or down (which it isn't, I am connected to it with browser right now). I have triple checked my stunnel scripts and everything should be ok. Could it like http://www.stunnel.org/faq/troubleshooting.html#ToC17? Any ideas? Markus ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] node on a server, and at home?
don't worry, your node will automatically tell you when a newer version i= s available (by noticing new build numbers within the network) Usually :). Check the web site from time to time, or read this list or devl, in case there's a reset. This is not necessarily so. 99% of the time this works, but the most recent node update (5090) was a full and clean network reset, meaning 5090 would talk only to builds 5090 and above. If you had a 5089 node, you'd never get the notice about seeing a newer build. This doesn't immediately matter to you, Mike, I'm just commenting on the above statement. true, you're both correct, in fact someone simply *has* to use the new builds first so the other builds can see there's something new around. maybe my answer was shot too fast again ^-^ the comment was meant as an easy small step for node ops when to update; being member of the support, announce and devl mailinglist is of course superior and preferable hand ;) ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] 5090 quirks...
ack.
then i hope you're pleased with my other observations?
so one can say, a node is unusable without doubt if it has less than, say, 40
connections to other nodes, is halfway useable if less than 60 and after that it's
okay?
maybe the mainport could give credit to this and not only show the 'first time user
message' but also a box describing verbally how many connections there are and what
successes the user can expect with that
amount of connections. this might hopefully dam up the 'help! my node can't connect'
mails and increase 1st time users' rating. communication is usually helpful ;)
especially between cryptic and strange nodes and
freenet maidens
It rejected the requests because it only had one connection. It had
nowhere to route the request to and didn't have it in its routing table,
so it instantly RNF'd it.
On Mon, Aug 09, 2004 at 10:54:22PM +0200, [EMAIL PROTECTED] wrote:
hi there, today i've tried the new 5090 build and i'd like to share my ex=
periences with you.
=20
first of, i've run into the 'too big seednodes' problem, too cutting =
the seednodes into half and throwing away one of these halves helped, as we=
ll as the suggestion to strip the file of every 'estimator' line, which=20
worked very well too (dunno if it breaks something either, but neverthele=
ss all the noderefs seem to show up in the RT)
=20
one of the first times i started my node, it had about ~30 connections to=
other nodes after not quite 1 hour. then i restarted the node (because i n=
eeded full network bandwidth). the next node start provided me=20
ONE connection to another node in the first 20 minutes. even after 2h of =
uptime i've come to only 20 live connections. that's weird and very depress=
ing :-/ as you can imagine the node was never really useable as=20
it was constantly backed off by all nodes it had connections to.
=20
what i've discovered then is the main reason why i write this mail.
at the time my node had this only one connection to the other node i was =
able to track the type of the messages which got passed between the two nod=
es.
interesting was, that the foreign node (i will now call it 'node B') was =
quite gentle to my node ('A') as it routed some DataRequests and later so=
me StoreDatas into my direction. so one can say that node B tried to=20
integrate my node into the network and thus began to route some things in=
to my direction. not too many, but what i'd like to call just right, mean=
s something like around 1 message per 1 minute. (hm, i suppose it=20
could be more)
=20
after some time the passed message types shown at the ocm connections pag=
e looked like this:
=20
Accepted 3/1=09
DataNotFound 0/1=09
QueryRejected3/0=09
DataRequest 1/3=09
=20
he send 3 DataRequests, i sent 3 Accepted, and now it comes.. my node res=
ponded immediately with 3 QueryRejecteds! (all numbers were always equal wh=
en reloading the page, 2=3D2=3D2, 4=3D4=3D4, ...)
the question is: why did my node reject the query?
see the following stats:
=20
Current routingTime 0ms=09
Current messageSendTimeRequest 0ms=09
Pooled threads running jobs 47 (39,2%)=09
Pooled threads which are idle7=09
Current upstream bandwidth usage 76 bytes/second (1,9%)=09
Current estimated load for QueryReject purposes 39%=09
Current estimated load for rate limiting 39,2%=09
Reason for load: Load due to thread limit =3D 39,2%
Load due to routingTime =3D 10% =3D 100ms / 1000ms =3D overloadLow (100%)
Load due to messageSendTimeRequest =3D 20% =3D 100ms / 500ms =3D overloa=
dLow=20
(100%)
Load due to output bandwidth limiting =3D 2,3% because outputBytes(4589) =
=3D=20
limit (196608,003 ) =3D outLimitCutoff (0,8) * outputBandwidthLimit (4096=
) *=20
60
Load due to expected inbound transfers: 0,5% because: 1000.0 req/hr *=20
9.950189371914758E-4 (pTransfer) * 86016.0 bytes =3D 85587 bytes/hr expec=
ted=20
from current requests, but maxInputBytes/minute =3D 245760 (set input lim=
it) *=20
60 * 1.1 =3D 16220160 bytes/hr target
Load due to expected outbound transfers: 4,2% because: 5046.5665649684115=
=20
req/hr * 9.970089730807576E-4(2 0s, 0 1s, 2 total) (pTransfer) * 86016.0=
=20
bytes =3D 432787 bytes/hr expected from current requests, but=20
maxInputBytes/minute =3D 172032 * 60 * 0.8 =3D 10321920 bytes/hr target=
=20
=20
my node was *never* overloaded --okay, the first few seconds after node =
startup the messageSTR was huge-- but after that the node was constantly a=
round 20-40% load; bw was nearly unused (i set bw to poor=20
4kb/s up and down, but that was even never reached. a later removal of th=
e limit did not help, either), cpu usage was very low too as nothing happen=
ed on the node.
=20
why should the node answer with a QR? i don't get it.
=20
of course the other node will be disppointed by my node's performance and=
decide to route somewhere else (or it ignores the QR and hammers regardles=
s earning even more
Re: [freenet-support] node on a server, and at home?
Ouch, typo in Fuqids configuration (wrong port). Now it doesn't say that node is overloaded or down, but I still can't get it to download anything through stunnel, logs say Fatal error in download thread: EFCPError: recv failed: 10054. Markus Hi I'm also running a node on my private serverfarm and accessing it from home. I succesfully created ssl tunnel with stunnel from my Linux router to my Linux server for surfing freenet with my browser. My problem is that I'm having trouble to create working stunnel between Fuqid (running on WinXP) and my nodes fcp-port, Fuqid tells me that my node is overloaded or down (which it isn't, I am connected to it with browser right now). I have triple checked my stunnel scripts and everything should be ok. Could it like http://www.stunnel.org/faq/troubleshooting.html#ToC17? Any ideas? Markus ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] node on a server, and at home?
On Tue, 10 Aug 2004 10:31:01 +0300 Markus [EMAIL PROTECTED] wrote: Ouch, typo in Fuqids configuration (wrong port). Now it doesn't say that node is overloaded or down, but I still can't get it to download anything through stunnel, logs say Fatal error in download thread: EFCPError: recv failed: 10054. 10054 is the Winsock error code for Connection Reset, which is likely indicative of a bad setup on the remote side. As Mika mentioned, first ensure that the remote node is allowing your IP address to connect to whichever service you're looking for - mainport or FCP. (For FUQID, it's FCP, and note that you lose anonymity making FCP requests to a node you aren't operating locally.) These are set via freenet.conf options on the server running the node. If everything checks out, FUQID isn't properly connecting to the remote node. My guess would be either that stunnel is not properly setup, or your remote node is so overloaded that it cannot accept your connection. First off, make sure that your FUQID is set up to use the remote node's IP address, or your local SSH tunnel to that remote IP. FUQID defaults to localhost on the default FCP port (8481). I am familiar with stunnel, I use it to initiate secured IRC sessions, but I haven't used it in conjunction with Freenet. What is the stunnel command you're using on each end? Feel free to XX.XX.XX.XX out IP addresses. Maybe I or someone else familiar with stunnel can help. -s ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
[freenet-support] Re: node on a server, and at home?
S writes: On Tue, 10 Aug 2004 10:31:01 +0300 Markus [EMAIL PROTECTED] wrote: Ouch, typo in Fuqids configuration (wrong port). Now it doesn't say that node is overloaded or down, but I still can't get it to download anything through stunnel, logs say Fatal error in download thread: EFCPError: recv failed: 10054. 10054 is the Winsock error code for Connection Reset, which is likely indicative of a bad setup on the remote side. As Mika mentioned, first ensure that the remote node is allowing your IP address to connect to whichever service you're looking for - mainport or FCP. (For FUQID, it's FCP, and note that you lose anonymity making FCP requests to a node you aren't operating locally.) These are set via freenet.conf options on the server running the node. One gotcha is that when you're using a tunnel, is that the node sees the IP address of the _endpoint_ of the tunnel, not the real source IP. So if you're tunneling straight to the computer running the node, access from localhost is enough. -- Mika Hirvonen [EMAIL PROTECTED] http://nightwatch.mine.nu/ Get Freenet from: http://cs181027153.pp.htv.fi:8891/J0~0J7ajDJE/ ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] node on a server, and at home?
Ouch, typo in Fuqids configuration (wrong port). Now it doesn't say that node is overloaded or down, but I still can't get it to download anything through stunnel, logs say Fatal error in download thread: EFCPError: recv failed: 10054. 10054 is the Winsock error code for Connection Reset, which is likely indicative of a bad setup on the remote side. As Mika mentioned, first ensure that the remote node is allowing your IP address to connect to whichever service you're looking for - mainport or FCP. (For FUQID, it's FCP, and note that you lose anonymity making FCP requests to a node you aren't operating locally.) These are set via freenet.conf options on the server running the node. Everything works fine without ssl tunnel (yes, I know, someone can sniff unencrypted connections). If everything checks out, FUQID isn't properly connecting to the remote node. My guess would be either that stunnel is not properly setup, or your remote node is so overloaded that it cannot accept your connection. First off, make sure that your FUQID is set up to use the remote node's IP address, or your local SSH tunnel to that remote IP. FUQID defaults to localhost on the default FCP port (8481). I am familiar with stunnel, I use it to initiate secured IRC sessions, but I haven't used it in conjunction with Freenet. What is the stunnel command you're using on each end? Feel free to XX.XX.XX.XX out IP addresses. Maybe I or someone else familiar with stunnel can help. -s I have following setup: LAN(192.168.1.0/24, client side, Fuqid) -- router (Linux router, firewall, runs stunnel, ip 192.168.1.1) -- INTERNET -- my.server.domainname (Linux server, firewall, runs stunnel, Freenet node, server side) I tried also with firewalls allowing all connections, didn't help. Server side stunnel: # For fproxy connection, this works fine stunnel -O l:TCP_NODELAY=1 -d 8880 -r localhost: # For FPC, does not work stunnel -O l:TCP_NODELAY=1 -d 8841 -r localhost:8481 client side stunnel: # Fproxy, works fine, http://192.168.1.1:8880 and https://my.server.domainname:8880 works stunnel -O r:TCP_NODELAY=1 -c -d 8880 -r my.server.domainname:8880 # FPC, does not work stunnel -O r:TCP_NODELAY=1 -c -d 8481 -r my.server.domainname:8841 Fuqid: unencrypted connections work IP: xxx.xxx.xxx.xxx port: 8481 encrypted connections fail IP: 192.168.1.1 port: 8841 Markus ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Re: node on a server, and at home?
S writes: On Tue, 10 Aug 2004 10:31:01 +0300 Markus [EMAIL PROTECTED] wrote: Ouch, typo in Fuqids configuration (wrong port). Now it doesn't say that node is overloaded or down, but I still can't get it to download anything through stunnel, logs say Fatal error in download thread: EFCPError: recv failed: 10054. 10054 is the Winsock error code for Connection Reset, which is likely indicative of a bad setup on the remote side. As Mika mentioned, first ensure that the remote node is allowing your IP address to connect to whichever service you're looking for - mainport or FCP. (For FUQID, it's FCP, and note that you lose anonymity making FCP requests to a node you aren't operating locally.) These are set via freenet.conf options on the server running the node. One gotcha is that when you're using a tunnel, is that the node sees the IP address of the _endpoint_ of the tunnel, not the real source IP. So if you're tunneling straight to the computer running the node, access from localhost is enough. -- Mika Hirvonen [EMAIL PROTECTED] http://nightwatch.mine.nu/ Get Freenet from: http://cs181027153.pp.htv.fi:8891/J0~0J7ajDJE/ Problem solved, it was that gotcha, FCP was dropping connections from localhost. I commented out fcpHosts=client.domain.name so Freenet allows localhost to connect fcp port. Markus ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
[freenet-support] My 5090 is dead
Windows XP Home 1.8 GHz 384 MB RAM Installed Build 5090 Load has been 0 to 1% for 3 days. Open Connections: Connections open (Inbound/Outbound/Limit) 0 (0/0/200) Transfers active (Transmitting/Receiving) 0 (0/0) Data waiting to be transferred None Total amount of data transferredNone Number of requests (sent/received) 0/0 Routing Table (at /nodestatus.html) Number of known routing nodes 0 Number of node references 0 Number of newbie nodes 0 Number of uncontactable nodes 0 Contacted and attempted to contact node references 0 Contacted node references 0 Contacted newbie node references0 Connections with Successful Transfers 0 Backed off nodes0 Connection Attempts 0 Successful Connections 0 Lowest max estimated search time0ms Lowest max estimated DNF time 0ms Lowest global search time estimate 30.0ms Highest global search time estimate 30.0ms Lowest global transfer rate estimate0 bytes/second Highest global transfer rate estimate 0 bytes/second Lowest one hop probability of DNF 0.95 Highest one hop probability of DNF 0.95 Lowest one hop probability of transfer failure 0.95 Highest one hop probability of transfer failure 0.95 Single hop probability of QueryRejected 0.9 Single hop average time for QueryRejected 17720.0 Single hop probability of early timeout 0.9 Single hop average time for early timeout 17720.0 Single hop probability of search timeout0.9 Single hop average time for search timeout 493544.0 Single hop overall probability of DNF given no timeout 0.9 Single hop overall probability of transfer failure given transfer 0.9 Probability of transfer given incoming request 0.1 Total number of requests that didn't QR 0 Total number of reqests that timed out before a QR or Accepted 0 Implementation freenet.node.rt.NGRoutingTable ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
[freenet-support] and null pointer exceptions...
Windows XP Home 1.8 GHz 384 MB RAM Installed Build 5090 Load has been 0 to 1% for 3 days. INFO: Native CPUID library 'freenet/support/CPUInformation/jcpuid-x86-windows.dll' loaded from resource INFO: Optimized native BigInteger library 'net/i2p/util/jbigi-windows-pentium4.dll' loaded from resource Aug 9, 2004 9:00:11 PM (freenet.client.http.FproxyServlet, YThread-0, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:16 PM (freenet.client.http.FproxyServlet, YThread-0, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:20 PM (freenet.client.http.FproxyServlet, YThread-4, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:20 PM (freenet.client.http.FproxyServlet, YThread-5, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:20 PM (freenet.client.http.FproxyServlet, YThread-7, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:20 PM (freenet.client.http.FproxyServlet, YThread-0, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:22 PM (freenet.client.http.FproxyServlet, YThread-3, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at
Re: [freenet-support] re: update.sh produces a 24Mb seednodes.ref?
Try changing it to -Xmx160M. That should fix the problem and allow you to reseed. You can change it back afterwards. 5091 will solve the problem by implementing a hack to use less memory, will be out today or tomorrow. Apologies for the delay. On Tue, Aug 10, 2004 at 04:07:50AM +0200, Heine Laursen wrote: Toad wrote: How do you invoke Freenet? Via start-freenet.sh ? What -Xmx does it use? Yes. [EMAIL PROTECTED]:~/freenet$ grep Xmx start-freenet.sh echo Command line: java -Xmx128m $JAVA_ARGS freenet.node.Main $@ nice -n 10 -- java -Xmx128m $JAVA_ARGS freenet.node.Main $@ [EMAIL PROTECTED]:~/freenet$ Sincerley Heine Laursen -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. signature.asc Description: Digital signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] 5090 quirks...
Ugh. No thanks, it's simply not that predictable, and it changes
constantly.
On Tue, Aug 10, 2004 at 09:11:51AM +0200, [EMAIL PROTECTED] wrote:
ack.
then i hope you're pleased with my other observations?
so one can say, a node is unusable without doubt if it has less than, say, 40
connections to other nodes, is halfway useable if less than 60 and after that it's
okay?
maybe the mainport could give credit to this and not only show the 'first time user
message' but also a box describing verbally how many connections there are and what
successes the user can expect with that
amount of connections. this might hopefully dam up the 'help! my node can't connect'
mails and increase 1st time users' rating. communication is usually helpful ;)
especially between cryptic and strange nodes and
freenet maidens
It rejected the requests because it only had one connection. It had
nowhere to route the request to and didn't have it in its routing table,
so it instantly RNF'd it.
On Mon, Aug 09, 2004 at 10:54:22PM +0200, [EMAIL PROTECTED] wrote:
hi there, today i've tried the new 5090 build and i'd like to share my ex=
periences with you.
=20
first of, i've run into the 'too big seednodes' problem, too cutting =
the seednodes into half and throwing away one of these halves helped, as we=
ll as the suggestion to strip the file of every 'estimator' line, which=20
worked very well too (dunno if it breaks something either, but neverthele=
ss all the noderefs seem to show up in the RT)
=20
one of the first times i started my node, it had about ~30 connections to=
other nodes after not quite 1 hour. then i restarted the node (because i n=
eeded full network bandwidth). the next node start provided me=20
ONE connection to another node in the first 20 minutes. even after 2h of =
uptime i've come to only 20 live connections. that's weird and very depress=
ing :-/ as you can imagine the node was never really useable as=20
it was constantly backed off by all nodes it had connections to.
=20
what i've discovered then is the main reason why i write this mail.
at the time my node had this only one connection to the other node i was =
able to track the type of the messages which got passed between the two nod=
es.
interesting was, that the foreign node (i will now call it 'node B') was =
quite gentle to my node ('A') as it routed some DataRequests and later so=
me StoreDatas into my direction. so one can say that node B tried to=20
integrate my node into the network and thus began to route some things in=
to my direction. not too many, but what i'd like to call just right, mean=
s something like around 1 message per 1 minute. (hm, i suppose it=20
could be more)
=20
after some time the passed message types shown at the ocm connections pag=
e looked like this:
=20
Accepted 3/1=09
DataNotFound 0/1=09
QueryRejected 3/0=09
DataRequest1/3=09
=20
he send 3 DataRequests, i sent 3 Accepted, and now it comes.. my node res=
ponded immediately with 3 QueryRejecteds! (all numbers were always equal wh=
en reloading the page, 2=3D2=3D2, 4=3D4=3D4, ...)
the question is: why did my node reject the query?
see the following stats:
=20
Current routingTime0ms=09
Current messageSendTimeRequest 0ms=09
Pooled threads running jobs47 (39,2%)=09
Pooled threads which are idle 7=09
Current upstream bandwidth usage 76 bytes/second (1,9%)=09
Current estimated load for QueryReject purposes39%=09
Current estimated load for rate limiting 39,2%=09
Reason for load: Load due to thread limit =3D 39,2%
Load due to routingTime =3D 10% =3D 100ms / 1000ms =3D overloadLow (100%)
Load due to messageSendTimeRequest =3D 20% =3D 100ms / 500ms =3D overloa=
dLow=20
(100%)
Load due to output bandwidth limiting =3D 2,3% because outputBytes(4589) =
=3D=20
limit (196608,003 ) =3D outLimitCutoff (0,8) * outputBandwidthLimit (4096=
) *=20
60
Load due to expected inbound transfers: 0,5% because: 1000.0 req/hr *=20
9.950189371914758E-4 (pTransfer) * 86016.0 bytes =3D 85587 bytes/hr expec=
ted=20
from current requests, but maxInputBytes/minute =3D 245760 (set input lim=
it) *=20
60 * 1.1 =3D 16220160 bytes/hr target
Load due to expected outbound transfers: 4,2% because: 5046.5665649684115=
=20
req/hr * 9.970089730807576E-4(2 0s, 0 1s, 2 total) (pTransfer) * 86016.0=
=20
bytes =3D 432787 bytes/hr expected from current requests, but=20
maxInputBytes/minute =3D 172032 * 60 * 0.8 =3D 10321920 bytes/hr target=
=20
=20
my node was *never* overloaded --okay, the first few seconds after node =
startup the messageSTR was huge-- but after that the node was constantly a=
round 20-40% load; bw was nearly unused (i set bw to poor=20
4kb/s up and down, but that was even never reached. a later removal of th=
e limit did not help, either), cpu usage was very low too as nothing happen=
ed on
Re: [freenet-support] node on a server, and at home?
Don't use stunnel, use ssh, and forward the ports. Stunnel just creates an HTTPS connection, right? On Tue, Aug 10, 2004 at 10:31:01AM +0300, Markus wrote: Ouch, typo in Fuqids configuration (wrong port). Now it doesn't say that node is overloaded or down, but I still can't get it to download anything through stunnel, logs say Fatal error in download thread: EFCPError: recv failed: 10054. Markus Hi I'm also running a node on my private serverfarm and accessing it from home. I succesfully created ssl tunnel with stunnel from my Linux router to my Linux server for surfing freenet with my browser. My problem is that I'm having trouble to create working stunnel between Fuqid (running on WinXP) and my nodes fcp-port, Fuqid tells me that my node is overloaded or down (which it isn't, I am connected to it with browser right now). I have triple checked my stunnel scripts and everything should be ok. Could it like http://www.stunnel.org/faq/troubleshooting.html#ToC17? Any ideas? Markus ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED] -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. signature.asc Description: Digital signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
[freenet-support] datastore size
Can it hurt the network if I make my datastore too big? ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] node on a server, and at home?
Don't use stunnel, use ssh, and forward the ports. Stunnel just creates an HTTPS connection, right? From Stunnel's man page (http://www.stunnel.org/faq/stunnel.html#description): The stunnel program is designed to work as SSL encryption wrapper between remote clients and local (inetd-startable) or remote servers. The concept is that having non-SSL aware daemons running on your system you can easily set them up to communicate with clients over secure SSL channels. Would ssh port forwarding be better method? Markus ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Startup times
Am Sonntag, 8. August 2004 13:48 schrieb Anonymous: ... yes, disk is in full thrash.. cpu maxed. other processes grinding to a halt. ... 64mb (best this mb will do) ... Sorry to say that, but just forget this for now. My node is P1/200 with 128MB of RAM (Gentoo Linux) and i had to choke it really hard to get it working more than 12 hours without restart. (Now it does, but not very powerful operation) It's not really the speed, the RAM matters. btw. if you could obtain 32MB PS2-Simms or even 64MB chips, you may give your computer more (i'm quite sure), but these chips are hard to get now, they are ancient in computer evolution terms. good byte pgpMShgg6t0Uh.pgp Description: signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] node on a server, and at home?
Ahh, okay. As long as it doesn't need the clients to be ssl aware. On Tue, Aug 10, 2004 at 10:48:44PM +0300, Markus wrote: Don't use stunnel, use ssh, and forward the ports. Stunnel just creates an HTTPS connection, right? From Stunnel's man page (http://www.stunnel.org/faq/stunnel.html#description): The stunnel program is designed to work as SSL encryption wrapper between remote clients and local (inetd-startable) or remote servers. The concept is that having non-SSL aware daemons running on your system you can easily set them up to communicate with clients over secure SSL channels. Would ssh port forwarding be better method? -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. signature.asc Description: Digital signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] My 5090 is dead
Have you tried reseeding? On Tue, Aug 10, 2004 at 01:44:55AM +, Sam wrote: Windows XP Home 1.8 GHz 384 MB RAM Installed Build 5090 Load has been 0 to 1% for 3 days. Open Connections: Connections open (Inbound/Outbound/Limit) 0 (0/0/200) Transfers active (Transmitting/Receiving) 0 (0/0) Data waiting to be transferredNone Total amount of data transferred None Number of requests (sent/received) 0/0 Routing Table (at /nodestatus.html) Number of known routing nodes 0 Number of node references 0 Number of newbie nodes0 Number of uncontactable nodes 0 Contacted and attempted to contact node references0 Contacted node references 0 Contacted newbie node references 0 Connections with Successful Transfers 0 Backed off nodes 0 Connection Attempts 0 Successful Connections0 Lowest max estimated search time 0ms Lowest max estimated DNF time 0ms Lowest global search time estimate30.0ms Highest global search time estimate 30.0ms Lowest global transfer rate estimate 0 bytes/second Highest global transfer rate estimate 0 bytes/second Lowest one hop probability of DNF 0.95 Highest one hop probability of DNF0.95 Lowest one hop probability of transfer failure0.95 Highest one hop probability of transfer failure 0.95 Single hop probability of QueryRejected 0.9 Single hop average time for QueryRejected 17720.0 Single hop probability of early timeout 0.9 Single hop average time for early timeout 17720.0 Single hop probability of search timeout 0.9 Single hop average time for search timeout493544.0 Single hop overall probability of DNF given no timeout0.9 Single hop overall probability of transfer failure given transfer 0.9 Probability of transfer given incoming request0.1 Total number of requests that didn't QR 0 Total number of reqests that timed out before a QR or Accepted0 Implementationfreenet.node.rt.NGRoutingTable ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED] -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. signature.asc Description: Digital signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] and null pointer exceptions...
Known bug will be fixed in 5091. On Tue, Aug 10, 2004 at 02:37:06AM +, Sam wrote: Windows XP Home 1.8 GHz 384 MB RAM Installed Build 5090 Load has been 0 to 1% for 3 days. INFO: Native CPUID library 'freenet/support/CPUInformation/jcpuid-x86-windows.dll' loaded from resource INFO: Optimized native BigInteger library 'net/i2p/util/jbigi-windows-pentium4.dll' loaded from resource Aug 9, 2004 9:00:11 PM (freenet.client.http.FproxyServlet, YThread-0, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:16 PM (freenet.client.http.FproxyServlet, YThread-0, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:20 PM (freenet.client.http.FproxyServlet, YThread-4, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:20 PM (freenet.client.http.FproxyServlet, YThread-5, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:20 PM (freenet.client.http.FproxyServlet, YThread-7, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:20 PM (freenet.client.http.FproxyServlet, YThread-0, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at javax.servlet.http.HttpServlet.service(Unknown Source) at freenet.interfaces.servlet.ServletContainer.handle(ServletContainer.java:82) at freenet.interfaces.LocalNIOInterface$ConnectionShell.run (LocalNIOInterface.java:268) at freenet.thread.YThreadFactory$YThread.run(YThreadFactory.java:285) Aug 9, 2004 9:00:22 PM (freenet.client.http.FproxyServlet, YThread-3, ERROR): Unexpected Exception in FproxyServlet.doGet -- java.lang.NullPointerException java.lang.NullPointerException at freenet.Key.size(Key.java:269) at freenet.client.http.FproxyServlet.doGet(FproxyServlet.java:568) at javax.servlet.http.HttpServlet.service(Unknown Source) at
[freenet-support] h
Am Dienstag, 10. August 2004 21:41 schrieb Steve: Can it hurt the network if I make my datastore too big? short answer: No! long answer: The store fills up when your node serves requests from others and you (the user). If a requested key is already in the store, your node just sends this and makes no request on his own for that. When the store is full, the least recently used key(s) is/are purged to make room. A bigger store means more keys are available, so chances to find requested keys in the store increase, this is _GOOD_ for the network. ( might only overflow your own hard disk if configured insane :-) ) good byte berny [EMAIL PROTECTED] pgph4g26X2fpg.pgp Description: signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
[freenet-support] oops
sorry, messed up the last mail a bit, but the text is valid! good byte pgpkft2zi46m6.pgp Description: signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] access freenet
Am Dienstag, 10. August 2004 20:23 schrieb Ralph Towner: Can you please help me. For years I've never had problems to connect to freenet, but since some days ago I can't anymore. What could be the problem? I include my logfile. Thanks! The seednodes.ref file is quite big nowadays. that is the reason for the OutOfMemory messages which do lobotomize your node before starting up. Please use an editor of your choice and cut the file down to only some few megabytes; then restart. this shall be fixed with stable version 5091, until then, use this method. good byte pgpPOnXJ6rxIc.pgp Description: signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] datastore size
On Tuesday 10 August 2004 21:41, Steve wrote: Can it hurt the network if I make my datastore too big? Not really, but enlarging your datastore all the time will prevent specialisation. This isn't real bad, but specialisation improves routing. If I were you I would let it specialise at one point. -- Michael A. Kuijn [EMAIL PROTECTED] () ascii ribbon campaign - against html mail /\- against microsoft attachments ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Startup times
On Wednesday 11 August 2004 00:12, Martin Scheffler wrote: My node is P1/200 with 128MB of RAM (Gentoo Linux) and i had to choke it really hard to get it working more than 12 hours without restart. (Now it does, but not very powerful operation) Holy bleep! My PIV 2.66 GHz with 256 MB RAM had similar results. Since I added 256 MB more everything is just fine. -- Michael A. Kuijn [EMAIL PROTECTED] () ascii ribbon campaign - against html mail /\- against microsoft attachments ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
RE: [freenet-support] datastore size
Hmmm. Am I wrong to think there probably is an optimal store size for each node? My thinking is that as the store grows, the node draws more requests, which at some point will exceed the node's ability to service them all in a timely manner due to resource limits (probably bandwidth). As the store grows larger, even more requests come in, causing the node to service a shrinking proportion of the total number of requests directed towards it, which has a negative impact on the network. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Martin Scheffler Sent: Tuesday, August 10, 2004 3:40 PM To: Steve; [EMAIL PROTECTED] Subject: [freenet-support] h Am Dienstag, 10. August 2004 21:41 schrieb Steve: Can it hurt the network if I make my datastore too big? short answer: No! long answer: The store fills up when your node serves requests from others and you (the user). If a requested key is already in the store, your node just sends this and makes no request on his own for that. When the store is full, the least recently used key(s) is/are purged to make room. A bigger store means more keys are available, so chances to find requested keys in the store increase, this is _GOOD_ for the network. ( might only overflow your own hard disk if configured insane :-) ) good byte berny [EMAIL PROTECTED] ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] datastore size
Am Mittwoch, 11. August 2004 00:55 schrieb Michael Kuijn: On Tuesday 10 August 2004 21:41, Steve wrote: Can it hurt the network if I make my datastore too big? Not really, but enlarging your datastore all the time will prevent specialisation. This isn't real bad, but specialisation improves routing. If I were you I would let it specialise at one point. Sorry, but forget that. In Freenet there is no thing like a one peak specialisation for more than one minute. And when the network can hold more data, it increases the lifetime of seldom requested keys. Specialization changes over time, anyway. The only drawback may be that your ubernode becomes popular over time because of more successful replies :-) (much connections increase request ability aswell) Routing does not get better when your node has to purge old keys on new requests - maybe they are requested just one minute later - more net load :-( good byte pgpty6uif4ja1.pgp Description: signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] datastore size
Am Mittwoch, 11. August 2004 01:06 schrieb Derek Ferguson: Hmmm. Am I wrong to think there probably is an optimal store size for each node? Sorry, yes. My thinking is that as the store grows, the node draws more requests, which at some point will exceed the node's ability to service them all in a timely manner due to resource limits (probably bandwidth). As the store grows larger, even more requests come in, causing the node to service a shrinking proportion of the total number of requests directed towards it, which has a negative impact on the network. The first is the request success rate (which makes your node attractive), another measure (request rate limiting) tells the other nodes to reduce the requests/time, to balance out all the peers with your avaiable bandwidth. A big store is good (even for yourself). good byte pgp3utfkDm14e.pgp Description: signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
[freenet-support] Stable build 5091
Stable build 5091 is now available. Please upgrade. To upgrade: On Windows, use the update option on the start menu, if it is there. On linux, stop the node, run update.sh, and start it. On any platform, stop the node, fetch http://freenetproject.org/snapshots/freenet-latest.jar and overwrite your existing freenet.jar with it, then restart the node. Major Changes: * Fix an NPE in fproxy when downloading KSKs. * Reduce memory usage when reseeding. * Save the overall rate limiting averages to disk, and load them on the next startup. This should smooth out the startup spike. * Slight improvement to load estimation. * Some changes to the distribution servlet (let it download the new freenet-ext.jar as a splitfile, untested). -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. signature.asc Description: Digital signature ___ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
