[freenet-support] Load management: Some interesting figures and next steps
The by-HTL stats include the average distance between the key and the ideal location for each HTL. Yesterday on my node, the closest point for CHKs is 0.0003 at HTL 13, and for SSKs it is 0.0041 at HTL 14. Note the extra zero! So we are nearly 12 times better at routing for SSKs than for CHKs! Another time (today), the ideal for SSKs is 0.0006 at HTL 15 and 13, but for CHKs it is 0.0063 at HTL 15. So the good news is we converge pretty quickly - the bad news is for CHKs we seem to not be able to get close to the target. IMHO this is because of my turning off some of the fairness between types logic. It was not possible to make it work with the new load management changes. There is a new version that will be deployed with the final new load management. But it also shows what is possible, and IMHO confirms my theory that misrouting as a result of mishandling of load is the main reason why performance is relatively poor. I doubt that reinstating the old fairness between types logic would do anything more than converge them both on 0.003 or thereabouts. What we need is the new load management logic. NEXT STEPS: * 1310 includes the first part of the bulk flag. This is necessary for new load management because new load management involves queueing i.e. is relatively high latency; by marking low latency requests explicitly, we can optimise them for latency, while using more of the available capacity for requests marked bulk that only care about throughput (most requests imho). Realtime requests have shorter timeouts, have priority for transfers (but with a scheme to prevent starvation), are accepted by load limiting in much fewer numbers, and are assumed to be very bursty. * Next week I will introduce code to actually set the bulk flag. For now it is always set to bulk. Next week we will set it to realtime for fproxy requests, and allow FCP apps to set it if they want. * We will also eliminate turtles, and increase the number of bulk requests allowed significantly. * Requests which are rejected due to short-term load management are marked as soft rejects. A very small change to make the node not remember such requests, so that other nodes can try again and maybe get accepted, should improve routing slightly. It is also a prerequisite for new-load-management which uses a limited number of retries in some cases (because while we know how many requests we are responsible for, we don't know how many requests other nodes are doing on our peer). * Two-stage timeout is important for new load management: If a downstream node times out, we need our peer to tell us that it has recognised the timeout. This allows us to know exactly when it is no longer running the request, and thus have an accurate count of how many of our requests are running on our peer, which is vital for new-load-management. * We will start sending the messages which indicate exactly how many requests are running and how many can be accepted from the peer. No more than one of these (for each of bulk and realtime) will be included in any given packet and it will be up to date at time of sending. * Then we can actually use them! The core of new load management is queueing so that we can get routed to a node reasonably close to our ideal. If a node is way below the median capacity, or if it is severely backed off, we won't wait for it, but otherwise we try to wait for our first choice. Lots of work has already been done towards this, for instance recently the changes making requests essentially threadless are a prerequisite for the bulk flag's greatly increasing the number running. signature.asc Description: This is a digitally signed message part. ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:support-requ...@freenetproject.org?subject=unsubscribe
[freenet-support] Load over 100 percent?
Today I've noticed that on my node's status page it says: Build 5107 ... Load: ... 137% It never went over 100 % before, so what does it mean? I did keep Frost running overnight...
[freenet-support] Load over 100 percent?
Today I've noticed that on my node's status page it says: Build 5107 ... Load: ... 137% It never went over 100 % before, so what does it mean? I did keep Frost running overnight... ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Load
David Masover wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Toad wrote: | No, if you can do that, then you can portscan for Freenet nodes. That's | a REALLY bad idea. You need to use some sort of seednodes mechanism. Why is that a bad idea? If a government is paranoid enough, they can just put devices all over which block any crypted traffic. Some sort of way for an effective seednodes file to be chosen or built quickly, then. |Obviously clients should be able to override that, but I think that fast |node connection is feasable, if The Network (TM) was fast enough. | | | Maybe so. But we want them to be USEFUL to the network. That won't | happen if they're only up for 5 minutes. They are useful if they make it more popular. Suppose there is some lag, say, 30 seconds to a minute to get an effective seednodes.ref. On top of that, such a node probably wouldn't be able to have a terribly big cache. Then there's incentive for people to whine for permanent nodes, and it wouldn't be too long before these Public Access Internet Terminals (or whatever hype word they use now) start being preloaded with Freenet. When governments gets involved in censorship, it is not usually that they are paranoid, but that they want the power to control. Sometimes to control the masses, but more often just to control the unloyal opposition, or those who simply disagree with them. ___ 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] Load
excuse me but seednodes looks a great point to quit the entire freenet, i start to use freenet because of it's working philosofy. my $0,02 On Thu, 2004-07-22 at 23:51, David Masover wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Toad wrote: | No, if you can do that, then you can portscan for Freenet nodes. That's | a REALLY bad idea. You need to use some sort of seednodes mechanism. Why is that a bad idea? If a government is paranoid enough, they can just put devices all over which block any crypted traffic. Some sort of way for an effective seednodes file to be chosen or built quickly, then. |Obviously clients should be able to override that, but I think that fast |node connection is feasable, if The Network (TM) was fast enough. | | | Maybe so. But we want them to be USEFUL to the network. That won't | happen if they're only up for 5 minutes. They are useful if they make it more popular. Suppose there is some lag, say, 30 seconds to a minute to get an effective seednodes.ref. On top of that, such a node probably wouldn't be able to have a terribly big cache. Then there's incentive for people to whine for permanent nodes, and it wouldn't be too long before these Public Access Internet Terminals (or whatever hype word they use now) start being preloaded with Freenet. -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQQB9SHgHNmZLgCUhAQI/iBAAlmwi6Cq3M9hsrkFPMzt7R1kZm2UmVhlo FBOWVJSfNQMxNsMYNOiwdPLHBqRIoIvKJsvIbTORGu9/M283c2Voqq14RhprL14n nA518/8cAmhoBrNJuMXg5ji8IqDbX4Jn4zg40HGf5nvxSsAveGmbfVZ2T7NveKZO 6Cs87GCuxlizgstOFRKP6S5Tsvrk16skQzyevwDm9tClbS5w2mbOFnk46h1uORQB kOPLP4vggS8f6QZPAZdsTU+iKZ3FHOo9ewY2zPWj43H6vYUkejIqLzas0/IZhSkL 6xkpJKT4tg33JAmrYiPp8c/gDBUoESP6F7OrbEccsc4TYb6z/mKWO2CZa/6m5XLQ z+D1W++uypAHr3KCPCg0ejHkIHHb3ixv58vWE4Ov+g5wyrio9eKxQyVNhHmegzkg v8ttcbXsrATnoNUv2qHbIPspqpJctbgE9+d/0dcwRmHII2465GeoZNW62slJ9goO 55WNgG9kpKQpmPaYok3zTiHsd2cTHxIEbRgEKG5SK2lx9d8SHqOuCnFWKJkY7LZZ Vgk//qqzN3dbsGDB821uhjmXHeBnNy1awjwy0EufuzA28G0tiRHY/dNLI+BUtBep beHP/K8h5wBWRIrknvtFYxKQY+WK2BrXFdcjB+EYnGv/r/ySxRbJ/6yURIMD9JaG sUgy81bqMtw= =a4R+ -END PGP 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] signature.asc Description: This is a digitally signed message part ___ 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] Load
I'm not sure exactly what you mean, but: In the short to medium term, central seednodes are a necessary evil. Especially as there is very little use of the Distribution Servlet/ Spread Freenet thing for more organic spreading of Freenet. Right now, Freenet traffic is relatively easy to detect. This will be fixed with Sessionv2 at some point before 1.0. In the longer term, IF routing works, it may be possible for Freenet to work with a more static mesh. Combined with steganography this would work reasonably well in hostile environments where seednodes would be dangerous. Although if the attacker has access to the whole network, they can probably do significant damage just by tracing node connections. But the point: seednodes are not necessarily the last word on all this. On Fri, Jul 23, 2004 at 10:38:33AM -0300, Lu?s Vit?rio Cargnini wrote: excuse me but seednodes looks a great point to quit the entire freenet, i start to use freenet because of it's working philosofy. my $0,02 On Thu, 2004-07-22 at 23:51, David Masover wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Toad wrote: | No, if you can do that, then you can portscan for Freenet nodes. That's | a REALLY bad idea. You need to use some sort of seednodes mechanism. Why is that a bad idea? If a government is paranoid enough, they can just put devices all over which block any crypted traffic. Some sort of way for an effective seednodes file to be chosen or built quickly, then. |Obviously clients should be able to override that, but I think that fast |node connection is feasable, if The Network (TM) was fast enough. | | | Maybe so. But we want them to be USEFUL to the network. That won't | happen if they're only up for 5 minutes. They are useful if they make it more popular. Suppose there is some lag, say, 30 seconds to a minute to get an effective seednodes.ref. On top of that, such a node probably wouldn't be able to have a terribly big cache. Then there's incentive for people to whine for permanent nodes, and it wouldn't be too long before these Public Access Internet Terminals (or whatever hype word they use now) start being preloaded with Freenet. -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQQB9SHgHNmZLgCUhAQI/iBAAlmwi6Cq3M9hsrkFPMzt7R1kZm2UmVhlo FBOWVJSfNQMxNsMYNOiwdPLHBqRIoIvKJsvIbTORGu9/M283c2Voqq14RhprL14n nA518/8cAmhoBrNJuMXg5ji8IqDbX4Jn4zg40HGf5nvxSsAveGmbfVZ2T7NveKZO 6Cs87GCuxlizgstOFRKP6S5Tsvrk16skQzyevwDm9tClbS5w2mbOFnk46h1uORQB kOPLP4vggS8f6QZPAZdsTU+iKZ3FHOo9ewY2zPWj43H6vYUkejIqLzas0/IZhSkL 6xkpJKT4tg33JAmrYiPp8c/gDBUoESP6F7OrbEccsc4TYb6z/mKWO2CZa/6m5XLQ z+D1W++uypAHr3KCPCg0ejHkIHHb3ixv58vWE4Ov+g5wyrio9eKxQyVNhHmegzkg v8ttcbXsrATnoNUv2qHbIPspqpJctbgE9+d/0dcwRmHII2465GeoZNW62slJ9goO 55WNgG9kpKQpmPaYok3zTiHsd2cTHxIEbRgEKG5SK2lx9d8SHqOuCnFWKJkY7LZZ Vgk//qqzN3dbsGDB821uhjmXHeBnNy1awjwy0EufuzA28G0tiRHY/dNLI+BUtBep beHP/K8h5wBWRIrknvtFYxKQY+WK2BrXFdcjB+EYnGv/r/ySxRbJ/6yURIMD9JaG sUgy81bqMtw= =a4R+ -END PGP 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] ___ 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] Load
On Wed, Jul 21, 2004 at 10:19:55PM -0500, David Masover wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Toad wrote: | java -cp freenet.jar freenet.Version on a command line? Freenet: Fred 0.5 (protocol STABLE-1.50) build 5084 (last good build: 5083) | Show me it. Same as what you've been seeing in the logs, pretty much. Since it dies now same as ever. Again, those urls are http://slaphack.com/freenet.conf http://slaphack.com/freenet.log | Thanks. That log appears to start from well into execution. Ah, no, it's | because of logLevel=error. Oops. logLevel=Normal, for now. | java -version java version 1.4.1 Java(TM) 2 Runtime Environment, Standard Edition (build Blackdown-1.4.1-01) Java HotSpot(TM) Client VM (build Blackdown-1.4.1-01, mixed mode) || Not possible. Java applets are not able to connect to servers other than | |Are not allowed to. They seem to be able to, if the user clicks yes |on a do you trust these people? dialog box. | | | Are they? Hmm. If you can prove that that would be really interesting. Googling for Java IRC Client... http://www.jpilot.com/products/jirc/demo1.html You're allowed to connect out to anywhere, if the user grants you permission. Or is it just me? Does this work with the bastardized Microsoft VM? Hrrm. Well I don't have a JVM on Mozilla, so I can't check... | However I don't see Freenet running well as a Java applet, because it | won't run for long enough... Freenet clients should be able to find nearby nodes fairly quickly. I mean, check which hosts on the local network have a Freenet port open, traceroute to figure out if we're on NAT and try to find nodes near the gateway, etc etc. No, if you can do that, then you can portscan for Freenet nodes. That's a REALLY bad idea. You need to use some sort of seednodes mechanism. Obviously clients should be able to override that, but I think that fast node connection is feasable, if The Network (TM) was fast enough. Maybe so. But we want them to be USEFUL to the network. That won't happen if they're only up for 5 minutes. -- 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] Load
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Toad wrote: | No, if you can do that, then you can portscan for Freenet nodes. That's | a REALLY bad idea. You need to use some sort of seednodes mechanism. Why is that a bad idea? If a government is paranoid enough, they can just put devices all over which block any crypted traffic. Some sort of way for an effective seednodes file to be chosen or built quickly, then. |Obviously clients should be able to override that, but I think that fast |node connection is feasable, if The Network (TM) was fast enough. | | | Maybe so. But we want them to be USEFUL to the network. That won't | happen if they're only up for 5 minutes. They are useful if they make it more popular. Suppose there is some lag, say, 30 seconds to a minute to get an effective seednodes.ref. On top of that, such a node probably wouldn't be able to have a terribly big cache. Then there's incentive for people to whine for permanent nodes, and it wouldn't be too long before these Public Access Internet Terminals (or whatever hype word they use now) start being preloaded with Freenet. -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQQB9SHgHNmZLgCUhAQI/iBAAlmwi6Cq3M9hsrkFPMzt7R1kZm2UmVhlo FBOWVJSfNQMxNsMYNOiwdPLHBqRIoIvKJsvIbTORGu9/M283c2Voqq14RhprL14n nA518/8cAmhoBrNJuMXg5ji8IqDbX4Jn4zg40HGf5nvxSsAveGmbfVZ2T7NveKZO 6Cs87GCuxlizgstOFRKP6S5Tsvrk16skQzyevwDm9tClbS5w2mbOFnk46h1uORQB kOPLP4vggS8f6QZPAZdsTU+iKZ3FHOo9ewY2zPWj43H6vYUkejIqLzas0/IZhSkL 6xkpJKT4tg33JAmrYiPp8c/gDBUoESP6F7OrbEccsc4TYb6z/mKWO2CZa/6m5XLQ z+D1W++uypAHr3KCPCg0ejHkIHHb3ixv58vWE4Ov+g5wyrio9eKxQyVNhHmegzkg v8ttcbXsrATnoNUv2qHbIPspqpJctbgE9+d/0dcwRmHII2465GeoZNW62slJ9goO 55WNgG9kpKQpmPaYok3zTiHsd2cTHxIEbRgEKG5SK2lx9d8SHqOuCnFWKJkY7LZZ Vgk//qqzN3dbsGDB821uhjmXHeBnNy1awjwy0EufuzA28G0tiRHY/dNLI+BUtBep beHP/K8h5wBWRIrknvtFYxKQY+WK2BrXFdcjB+EYnGv/r/ySxRbJ/6yURIMD9JaG sUgy81bqMtw= =a4R+ -END PGP 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] Load
David Masover wrote: Of course, if you don't own your own computer, how can you trust it? One-way trust. Suppose my bro trusts me, but I don't trust him, I have root, and he wants Freenet. You don't need root to run it and it's probably a good idea to not run it as root even when you are root. # useradd -r -d /path/to/freenet freenet # su - freenet -c /path/to/start-freenet.sh Z -- Framtiden är som en babianröv, färggrann och full av skit. Arne Anka ___ 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] Load
On Tue, Jul 20, 2004 at 11:48:29PM -0500, David Masover wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Toad wrote: | On Tue, Jul 20, 2004 at 01:33:30AM -0500, David Masover wrote: | |Unfortunately, I can't work on this at all right now. My freenet node |looks fine, only I get a connection close from FProxy the instant I try |connecting -- that is, 0 bytes sent/recieved from netcat, The document |contains no data from Firefox. | | | Ouch. Anything in the logs? Tried restarting? What build? Only about 20 times, how do I check the build without a working FProxy? java -cp freenet.jar freenet.Version on a command line? Trying to update to latest stable: Usually in the logs, it at least tells me something like starting Freenet, Build X Now, I get a java.io.IOException as the first log message. Show me it. Takes more than 3 mins for port to start listening. When it does, wget reports connection reset by peer. Deleting everything in /var/freenet except seednodes.ref, restarting... It's been 5 minutes. Log (/var/freenet/freenet.log) is still empty. Still not responding. Some time later, I check back, and -- yet again -- Connection reset by peer. For debugging purposes (I'm smart enough to remove this once I get freenet working), I'm going to leave this atrocity running overnight, with logs and conf files online: http://slaphack.com/freenet.log Thanks. That log appears to start from well into execution. Ah, no, it's because of logLevel=error. Hmm. This is a result of Yet Another 1.4.2 Big* Bug. I suggest you upgrade to 1.4.2-r05 or 1.5.0-beta2, or downgrade to 1.4.1. What build of the JVM are you running? You can find out by running the following on a command line: java -version http://slaphack.com/freenet.conf | || of the user. If a non-anonymous search solves one part without || affecting the other, what's the harm of it? | |None, as long as it's _absolutely_clear_ which parts are solved. If you |make the publisher anonymous but the readers known, you don't want |someone saying ooh, freenet and then using it to visit Porn of Love |from an office computer. Or worse. Make huge, bold, red warnings. | | | Hehe. That's their own silly fault ;). But yes, warnings probably a good | idea. Yes, and yes. n00bishness should be no barrier to revolution. | Google keeps the entire index in RAM. *jaw drops* You learn something new every day. |[...] || to reach the users of freenet and hushmail; I want it to || reach the huge and clueless masses who watch CNN and use || hotmail. And I also want to protect my anonymity damn well. | |What about implementing a freenet client as a Java applet, thus allowing |freenet gateways? You'd need that (no _percieved_ download/install of |software) in order to reach these huge and clueless masses. | | | Not possible. Java applets are not able to connect to servers other than Are not allowed to. They seem to be able to, if the user clicks yes on a do you trust these people? dialog box. Are they? Hmm. If you can prove that that would be really interesting. However I don't see Freenet running well as a Java applet, because it won't run for long enough... | running a public proxy. However tens of millions of lusers actually know | how to install software. Sometimes the problem is knowing how. Sometimes it's being allowed to. Not everyone owns their own computer. Of course, if you don't own your own computer, how can you trust it? One-way trust. Suppose my bro trusts me, but I don't trust him, I have root, and he wants Freenet. Or suppose someone doesn't care about their own anonymity, but wants Freenet on a public internet terminal. The possibilities are endless... Most end-user types I know don't ever upgrade their software as long as it seems to work, and few upgrade even when it stops working -- they reinstall. Web interface solves all of that, as long as browser cache / web proxies behave themselves. -- 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] Load
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Zenon Panoussis wrote: | | | David Masover wrote: | | Of course, if you don't own | your own computer, how can you trust it? One-way trust. Suppose my bro | trusts me, but I don't trust him, I have root, and he wants Freenet. | | | You don't need root to run it and it's probably a good idea | to not run it as root even when you are root. | # useradd -r -d /path/to/freenet freenet | # su - freenet -c /path/to/start-freenet.sh | Just going by the Gentoo package, but thanks, I'll keep that in mind. I'd rather make it work first and deal with that later. -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQP8pCHgHNmZLgCUhAQJmiBAAnY0vFcpdQzobqug8B9/F9jw4jxa0cBWT wEml2qCctZZkroolbFW4n2TBg5odm4po8MH2E0+thlIcJfdIjxttRFmTmcRZ82ld UZwdKPT6uwJqBbdsJ1m+8FN305KvKhXswWWwY/IlDsQRO0jkG24JTCwmKggGpm7X txPFK5VxY1HqiJ1xvgU6py9Y3qNmV+c0s+9XcHi+BDVpQBsxYpzM4abr0OoJMp9d NGcUZf13/ORuhaMAxTrr+rzXnESxbJCMj3W9kAy4xb2p5ExJpN8fzDgWyF2GC1Dc zDl6ZBiIDK7aOV8SSBi0ZU75yS6OIz784Ci2yz6tfp/2lni7J2xLO16/5tZj2yaY p1gq5Znjf08jtK43DZqICcSNae81whUuFfIEPpqTIYA3uivA0itAVGg+0mLCKZTC Ik3YH9/XX+FwR6eAdPUNbcPGMWotnl/6CBcA3E7Y+y4YwSWNPC4BxwiVmiti8Eqz wOFjaFp0Lv9b+goMLyJwRgaXV4Qd0ij+wptPiRz6vcbShuBecLz2qEPzFE/O3Q07 nRj7JD3e5PqPOS1KGHiC+GfCzOFg7jTmUF/k8qwBiFlp+Tip3WMboc5j/xOGxt5a C47Ri0rYbvOts3XdANMebq72YDqCUc+Qwx3ay8izuegl8tnZEkgCwAZcVmFUZXgD VRo1nWPJs9Y= =exCq -END PGP 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] Load
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Toad wrote: | java -cp freenet.jar freenet.Version on a command line? Freenet: Fred 0.5 (protocol STABLE-1.50) build 5084 (last good build: 5083) | Show me it. Same as what you've been seeing in the logs, pretty much. Since it dies now same as ever. Again, those urls are http://slaphack.com/freenet.conf http://slaphack.com/freenet.log | Thanks. That log appears to start from well into execution. Ah, no, it's | because of logLevel=error. Oops. logLevel=Normal, for now. | java -version java version 1.4.1 Java(TM) 2 Runtime Environment, Standard Edition (build Blackdown-1.4.1-01) Java HotSpot(TM) Client VM (build Blackdown-1.4.1-01, mixed mode) || Not possible. Java applets are not able to connect to servers other than | |Are not allowed to. They seem to be able to, if the user clicks yes |on a do you trust these people? dialog box. | | | Are they? Hmm. If you can prove that that would be really interesting. Googling for Java IRC Client... http://www.jpilot.com/products/jirc/demo1.html You're allowed to connect out to anywhere, if the user grants you permission. Or is it just me? Does this work with the bastardized Microsoft VM? | However I don't see Freenet running well as a Java applet, because it | won't run for long enough... Freenet clients should be able to find nearby nodes fairly quickly. I mean, check which hosts on the local network have a Freenet port open, traceroute to figure out if we're on NAT and try to find nodes near the gateway, etc etc. Obviously clients should be able to override that, but I think that fast node connection is feasable, if The Network (TM) was fast enough. -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQP8yW3gHNmZLgCUhAQJhZA/+JgN/uytQo5n1WKmPel2Y9LDM5V3096Je Sc/glsFEg512XBS/Gu/M2fdkPrn0pWPtiIbyAV4B6rBWwjhrOXiBe7bgRh8TV40X WP6OhF3wHqsTZImlgOsrCix/LUST/6l1qnsdHDjxkfNMaKWza9nUsIyzuvDLl5Ju mXGajo0/EzU8VYZ6rreY97s9BZtLTkcXJ++7m+XuNhrjiAQ64g9dimk43KF176T0 QOvSc4U8kFmKh4Oeb03wkWF1O+UoYl4E9OTLS16qVmOyDomF/wS2+SglUWR12XFl wm74jocGAZDBPE3FaDr5rb6AQ2gEWDJgHKZtZNJNd3ofhsq7l7AFtY7tJHRfrfy/ YhG+CNLbqpMHhHaPCms0IFk9lswaT2TLtbbJ2iCSL4qsn5jQXdXHkrcOQbcOhBI3 CGh17cKDRIDoIZHx54iDKDF+QWD4KAmv5URGzrF44ZgF/2aPQs9JQW51pYtI5Pf5 5TNQIwK3mFCgGOYaT5QGZAaZmy8bJzVRis3g9sK7hMq+bKFkF4kldawoWGv9ohsn Z+WPCxjsCgwkBKMWJ0SFd7pL+TTeSwTuk8pzAhnRA51gfN5LVHvOYzfu1hi68Mzd yikF5aJveXJuTN+HnC+NRDEAl0oOHGCH7hOYq+0Z0zEiXPPQtk0Ut+sxWn9Dwq12 X8i4S1vf9zg= =1w22 -END PGP 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] Load
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 TLD wrote: | what's your mainport.allowedHosts= setting, and what's the IP of the | computer you're trying to access from? It's sane. Or you tell me: http://slaphack.com/freenet.conf http://slaphack.com/freenet.log -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQP8yuHgHNmZLgCUhAQI0zQ/+POC5/kUI1lSIJuFRBN+D3zIRVBZBy0VE NTpouOBnnh5NSdoikl0Dr2I1wReyM6PVkCg7ToTcyZ66xhQ03wkDdLOTJ4IM73yY 7N7oI+gGl4B4emlm1oBZ1AExQVYGHrcZ98QU3+GdZo3MdROty5Vj3mRUhxz1KLDm 4673mNGEi4rtmL0ZAwONr7FTcf2QHF9HH/vwrYbG9VOE758Z6eNBuxBbtrFUrLnP fOq7tTFHqxaKZkRcmaSN0YMBjQC416xLicB90Grg9O6auqGQae57EJQmmveqCN8u eYFil0/JG6/46o0wA6eoXPkdIxX0eHztgPqiHgYSEfGOnBxkouwqRO5KKVdb6kMF HMDSa4BGp2MsBN420TkHy0z3/47KqY/6LyPkhr8SgBy3RJvYR5vI3GXG1eW3RoLc QU5HEYD3hH0fH8B2tiAMUAVPV3atpP3GJ22sqsgzNFyM710S80WiLW0RXEQpDl9/ EhmJDiZwyJCpCTBCeshweRKtKI0R7S5QFz0oBoxZPv0PyGydx+NiymVeO0IGqwAl ZrE1505ec+RfMVpzNoRK4/eBiZqCN08VMoA+GW//cXjhjIagYORMmILllf0xxQLs db7fp64Ph0ky/a0op6I+niIk12mu8SRgrYPdswMNcQP20o9AWOWu5RsLXlQo74VM Qrw1EkiMeWU= =Exvm -END PGP 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] Load
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Zenon Panoussis wrote: | | Toad wrote: | | The thing is, the lack of search capabilities reduces | the useability of freenet | | | Of course. There are ways to implement search, however. Sooner or later | somebody will implement a good spider based anonymous search. Sooner. NIM/FROST allows you to send information to someone. So the author of a new freesite uses one of those methods to notify the spider about their new freesite. Then the spider generates index files, and provides three interfaces for searching: - - Download the index files and a search client and run searches locally. With a fast freenet and small enough download chunks, this could be very effective -- at least until the index files get too big. - - Search via a real web interface -- a gateway somewhere. This compromises anonymity of the user, but is faster. - - Search by entering a query in NIM/FROST, and wait a bit for a response. This would require some sort of client software and a fast freenet to be sane. I read somewhere (probably on a freesite) about some sort of IRC which claimed to be completely anonymous. If there truly is something as fast as IRC and as anonymous as Freenet, it'd be very helpful for that third solution. Unfortunately, I can't work on this at all right now. My freenet node looks fine, only I get a connection close from FProxy the instant I try connecting -- that is, 0 bytes sent/recieved from netcat, The document contains no data from Firefox. | of the user. If a non-anonymous search solves one part without | affecting the other, what's the harm of it? None, as long as it's _absolutely_clear_ which parts are solved. If you make the publisher anonymous but the readers known, you don't want someone saying ooh, freenet and then using it to visit Porn of Love from an office computer. Or worse. Make huge, bold, red warnings. | This would | probably have two components: 1. A spider, which would spider out from | known freesites, scan NIMs, and Frost traffic, and insert index files. | 2. A client, probably integrated into fproxy, which would fetch the | index files that are appropriate to the search given. | | You mean creating index files before a search has been made? | Wouldn't that be highly inaccurate and/or produce massive | volumes of indices? It's what Google does. The spider, known as Googlebot, attempts to download the entire Internet and index it. When you run a Google search, it searches on that (albeit massive) index file. But the Google indices are probably huge. For a sample, download the DMOZ rdf dump -- http://rdf.dmoz.org/. [...] | to reach the users of freenet and hushmail; I want it to | reach the huge and clueless masses who watch CNN and use | hotmail. And I also want to protect my anonymity damn well. What about implementing a freenet client as a Java applet, thus allowing freenet gateways? You'd need that (no _percieved_ download/install of software) in order to reach these huge and clueless masses. You could just make a public FProxy, but then governments could ask you to give them the IPs of everyone who viewed a particular Freesite. Not good. -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQPy8uXgHNmZLgCUhAQLSBQ//evZOf0oDhszKotK0OTZI4qHPKuyCH9yJ njqC689ismMFbE8s/sK83XnHRwkn+4654jpwdzWvtKNjuHKtqhK/QnonbuGOJKvU I5fVyNVTMDSrtS22bzXRuZF6BIHGEA+jAihPktBA1uPTyfq5/pJhv1V8fkO41Uyw 7HvV7axRf56Zt9jgUjL410KH5/yeAIyumNJjv1dq/U7nSUNb//x5pAtFa/dim0xN clqGriaX1SiHMJsykDBWi99eBRBCyIgLktcsg9zMqZy3qP44EH/NeeSBFxsH/bte 0m2EXU9R3eYAUnGdX+dWN8ZAXBrxHtwtBK5MUniK4wlX0YSRy4cLdiMKkyvztCOs oGztyRCKDiFdOlr73vIOjcyIz7+JNhDqKYpfZzn2R/u93P2UxXLLaVL9Ael/dNxJ LA/DRbMce//skf+RgSQotfSg652zJHQIvsgpSZ/r8PPS9ReEpFbVEo5MzbclajCf R8SRf2kV+Qfq6oNzdK62D6JBO1B/35n4iUOSNt/asE6VQPxxjWlMRfc+JOhT/Lvz F3KvQ5ZqC3JMgMlwmyZ/6pB05dgrVriep0qhHWz+6+8t5Z5LnJFAhnEckgTE18sG 38tmIZxN57PLiUDS9zIaU8rzILZYulELwO+Z96RUV6XLcGJ5TrFnS3L0b9e0kBOs npRjlNmM/SA= =GUjB -END PGP 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] Load
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Zenon Panoussis Sent: den 20 juli 2004 05:15 To: [EMAIL PROTECTED] Subject: Re: [freenet-support] Load I wrote: Taking what you say here for granted, the entire discussion up to this point is probably a meaningless exchange based on some misunderstanding on my part. But what? [URIs from logs] Would be interested to see some of this list. Duh. So am I by now, but with all the messing around today I deleted them. I can try again though. Now I know what the misunderstanding was. The working URIs I found in my logs come from the default bookmarks in the interface servlet. I had never visited them before, but they had passed my client anyway. Yes, the node tries to preload them when it starts. Cheers /N ___ 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] Load
On Mon, Jul 19, 2004 at 11:50:02PM +0200, Zenon Panoussis wrote: Indeed. Thus we have NIMs, FreeMail and Frost within Freenet, and outside it we have Mixmaster remailers, IIP, I2P, various kinds of proxies and so on. Sadly some people use hushmail too, which is not exactly the safest option. But there are many possibilities. All this put together is still a *very* small world. If I'd find and publish, say, the Bush administration's plans to invade Cuba, or detailed information on Israel's chemical and biological weapons, I don't want this information to to reach the users of freenet and hushmail; I want it to reach the huge and clueless masses who watch CNN and use hotmail. And I also want to protect my anonymity damn well. The way to go? Publish on freenet and let automation, i.e. nobody, make the bridge to the web. Freenet is pretty small. I'd hijack an unsecured WEP in the middle of nowhere, and send the plans via a chain of remailers, direct to CNN. How do you propose to protect against spam, and plain malicious content? I don't. I'm not Google. As you have already gathered, my financial capacity is enough to run a 39-euro server, but not a 78-euro one. Because of that, things get very simple: if I make a freenet search, it will be just as well or ill protected from spam and malicious content as freenet itself is. In other words it will be spammed to hell in order to deny service. Freenet does not know the URIs of data that passes through the node, only those requested locally. It does know the requests that pass through the node. Nope. It doesn't. It only knows the routing keys, which are insufficient to decrypt the actual data. Any other URIs in the logs will be locally originated. Example: CHK@routing key,decrypt key/human readable key Uhm, there's something eluding me here. You know freenet's internals; I don't. If you say so, then so it is. Yet I stuck some of those URIs I found in my logs into my browser and got sites to which I had never been before. Browser prefetch? ARK insertions? A few other options. Taking what you say here for granted, the entire discussion up to this point is probably a meaningless exchange based on some misunderstanding on my part. But what? No, just run a spider. [URIs from logs] Would be interested to see some of this list. Duh. So am I by now, but with all the messing around today I deleted them. I can try again though. Are you running a public gateway? Are you fetching lots of stuff locally? Neither. -- 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] Load
On Tue, Jul 20, 2004 at 05:14:49AM +0200, Zenon Panoussis wrote: I wrote: Taking what you say here for granted, the entire discussion up to this point is probably a meaningless exchange based on some misunderstanding on my part. But what? [URIs from logs] Would be interested to see some of this list. Duh. So am I by now, but with all the messing around today I deleted them. I can try again though. Now I know what the misunderstanding was. The working URIs I found in my logs come from the default bookmarks in the interface servlet. I had never visited them before, but they had passed my client anyway. LOL! -- 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] Load
On Tue, Jul 20, 2004 at 01:33:30AM -0500, David Masover wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Zenon Panoussis wrote: | | Toad wrote: | | The thing is, the lack of search capabilities reduces | the useability of freenet | | | Of course. There are ways to implement search, however. Sooner or later | somebody will implement a good spider based anonymous search. Sooner. NIM/FROST allows you to send information to someone. So the author of a new freesite uses one of those methods to notify the spider about their new freesite. Then the spider generates index files, and provides three interfaces for searching: - - Download the index files and a search client and run searches locally. With a fast freenet and small enough download chunks, this could be very effective -- at least until the index files get too big. - - Search via a real web interface -- a gateway somewhere. This compromises anonymity of the user, but is faster. - - Search by entering a query in NIM/FROST, and wait a bit for a response. This would require some sort of client software and a fast freenet to be sane. I read somewhere (probably on a freesite) about some sort of IRC which claimed to be completely anonymous. If there truly is something as fast as IRC and as anonymous as Freenet, it'd be very helpful for that third solution. IIP, I2P, others. IIP died, but has been resurrected. Doesn't have enough users and relays to be really anonymous at the moment. Unfortunately, I can't work on this at all right now. My freenet node looks fine, only I get a connection close from FProxy the instant I try connecting -- that is, 0 bytes sent/recieved from netcat, The document contains no data from Firefox. Ouch. Anything in the logs? Tried restarting? What build? | of the user. If a non-anonymous search solves one part without | affecting the other, what's the harm of it? None, as long as it's _absolutely_clear_ which parts are solved. If you make the publisher anonymous but the readers known, you don't want someone saying ooh, freenet and then using it to visit Porn of Love from an office computer. Or worse. Make huge, bold, red warnings. Hehe. That's their own silly fault ;). But yes, warnings probably a good idea. | This would | probably have two components: 1. A spider, which would spider out from | known freesites, scan NIMs, and Frost traffic, and insert index files. | 2. A client, probably integrated into fproxy, which would fetch the | index files that are appropriate to the search given. | | You mean creating index files before a search has been made? | Wouldn't that be highly inaccurate and/or produce massive | volumes of indices? It's what Google does. The spider, known as Googlebot, attempts to download the entire Internet and index it. When you run a Google search, it searches on that (albeit massive) index file. But the Google indices are probably huge. For a sample, download the DMOZ rdf dump -- http://rdf.dmoz.org/. Google keeps the entire index in RAM. [...] | to reach the users of freenet and hushmail; I want it to | reach the huge and clueless masses who watch CNN and use | hotmail. And I also want to protect my anonymity damn well. What about implementing a freenet client as a Java applet, thus allowing freenet gateways? You'd need that (no _percieved_ download/install of software) in order to reach these huge and clueless masses. Not possible. Java applets are not able to connect to servers other than the one they were downloaded from. Thus you'd still be effectively running a public proxy. However tens of millions of lusers actually know how to install software. You could just make a public FProxy, but then governments could ask you to give them the IPs of everyone who viewed a particular Freesite. Not good. -- 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] Load
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Toad wrote: | On Tue, Jul 20, 2004 at 01:33:30AM -0500, David Masover wrote: | |Unfortunately, I can't work on this at all right now. My freenet node |looks fine, only I get a connection close from FProxy the instant I try |connecting -- that is, 0 bytes sent/recieved from netcat, The document |contains no data from Firefox. | | | Ouch. Anything in the logs? Tried restarting? What build? Only about 20 times, how do I check the build without a working FProxy? Trying to update to latest stable: Usually in the logs, it at least tells me something like starting Freenet, Build X Now, I get a java.io.IOException as the first log message. Takes more than 3 mins for port to start listening. When it does, wget reports connection reset by peer. Deleting everything in /var/freenet except seednodes.ref, restarting... It's been 5 minutes. Log (/var/freenet/freenet.log) is still empty. Still not responding. Some time later, I check back, and -- yet again -- Connection reset by peer. For debugging purposes (I'm smart enough to remove this once I get freenet working), I'm going to leave this atrocity running overnight, with logs and conf files online: http://slaphack.com/freenet.log http://slaphack.com/freenet.conf | || of the user. If a non-anonymous search solves one part without || affecting the other, what's the harm of it? | |None, as long as it's _absolutely_clear_ which parts are solved. If you |make the publisher anonymous but the readers known, you don't want |someone saying ooh, freenet and then using it to visit Porn of Love |from an office computer. Or worse. Make huge, bold, red warnings. | | | Hehe. That's their own silly fault ;). But yes, warnings probably a good | idea. Yes, and yes. n00bishness should be no barrier to revolution. | Google keeps the entire index in RAM. *jaw drops* You learn something new every day. |[...] || to reach the users of freenet and hushmail; I want it to || reach the huge and clueless masses who watch CNN and use || hotmail. And I also want to protect my anonymity damn well. | |What about implementing a freenet client as a Java applet, thus allowing |freenet gateways? You'd need that (no _percieved_ download/install of |software) in order to reach these huge and clueless masses. | | | Not possible. Java applets are not able to connect to servers other than Are not allowed to. They seem to be able to, if the user clicks yes on a do you trust these people? dialog box. | running a public proxy. However tens of millions of lusers actually know | how to install software. Sometimes the problem is knowing how. Sometimes it's being allowed to. Not everyone owns their own computer. Of course, if you don't own your own computer, how can you trust it? One-way trust. Suppose my bro trusts me, but I don't trust him, I have root, and he wants Freenet. Or suppose someone doesn't care about their own anonymity, but wants Freenet on a public internet terminal. The possibilities are endless... Most end-user types I know don't ever upgrade their software as long as it seems to work, and few upgrade even when it stops working -- they reinstall. Web interface solves all of that, as long as browser cache / web proxies behave themselves. -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQP31nHgHNmZLgCUhAQKsOQ//XDM6D434bGafKQmojJGqqq6QiQ5TZ49o EYfNz2HV3n/CSzD4llyy86HmflIwYqjs5M7EHe1iAGpkL13XQY8jj9NmtAU2gnk9 fEfMVoF0+Nx4odaXeopivXXqdYm7iSajrK4Au88P61/zfa/cu8xNdMnGQKtN33BZ +HrgWmm6uhfxTHOEj+GmH0Lav4C2I52LrQSoChabnzVJbUf9ejETV23VKJ2nrOZU wirBp2zeHwdiMDkmTBV4RAvn79Ae57m4IShhFLRjqlxxcgE2f2Fr6nJHRTDplhPX NYx5n4+VYbiRtqh9IsTrxEJDEu6yM2SeUdumb1N4fRCA0qPuAyMcd3olzinYP7df NiUKP+AtP3aruDs5yjTjIndn93QqmI+ek29+E/kUNySlYGbembqUbWlq0I+hwztk d2sYlOGPixU1/IBk4T926Lfj8X44XYBYm4nKYqUrFoSSM8ZpZhtxY8JO9u7bqiWw Qzn38ABcdQZoO6djBkEqBWgDRhBcd+rg9D3rOO5+RMmnrrLZoqcT++2AYUmdVzLn V6YQf1ZYmaqRZ8taglXf6+VW14UtlL3Fxp55BUofNG/1bVnu4WrxWwZJOWbPPotw Ehj52l+QKvV/tlRrvniGzEujS7BsaNzKsXUObmj8JmZq1XgKKKGh4x9oDQ/Y7SGI DVFVxgo3GJI= =JWYP -END PGP 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] Load
Hello everyone. I started a node on a machine with lots of bandwidth and a very lousy I/O subsystem. Not much else is going on on the machine, so without freenet the load is steadily between 0.01 and 0.10. When freenet runs, the load is constantly around 3.50, with peaks reaching well above 5.00. The system latency caused by these loads gets other stuff to malfunction; for instance, mail server queries to LDAP time out and result in service temporarily unavailable errors. Obviously, such things become a show stopper for freenet. The machine is a Celeron 2.4 GHz with 512 MB RAM running RHEL3 with Sun java 1.4.2_05. I reduced maxNodeConnections to 85 from the default 200 in the hope to reduce the number of java instances, but that didn't help much. I also set diagnosticsPath=/dev/null to stop the constant writing on disk (is there a better way to say disable stats, I don't need them?) and that didn't help much either. So I'm asking for advice from those more experienced with freenet: what can I do to reduce the load? More specifically, how can I reduce the number of java instances running? With my current settings (most else at default), I have 96 java processes eating away everything on the machine and the machine itself too. Z -- Framtiden r som en babianrv, frggrann och full av skit. Arne Anka ___ 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] Load
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Monday 19 July 2004 15:14, Zenon Panoussis wrote: Hello everyone. I started a node on a machine with lots of bandwidth and a very lousy I/O subsystem. Not much else is going on on the machine, so without freenet the load is steadily between 0.01 and 0.10. When freenet runs, the load is constantly around 3.50, with peaks reaching well above 5.00. The system latency caused by these loads gets other stuff to malfunction; for instance, mail server queries to LDAP time out and result in service temporarily unavailable errors. Obviously, such things become a show stopper for freenet. The machine is a Celeron 2.4 GHz with 512 MB RAM running RHEL3 with Sun java 1.4.2_05. I reduced maxNodeConnections to 85 from the default 200 in the hope to reduce the number of java instances, but that didn't help much. I also set diagnosticsPath=/dev/null to stop the constant writing on disk (is there a better way to say disable stats, I don't need them?) and that didn't help much either. So I'm asking for advice from those more experienced with freenet: what can I do to reduce the load? More specifically, how can I reduce the number of java instances running? With my current settings (most else at default), I have 96 java processes eating away everything on the machine and the machine itself too. I run freenet niced at +10 on a 2x500MHz computer, load stays at 2-3 all the time. I suspect the problem you have lies in the fact that freenet will eat ALL available bandwidth that you give it, which will lead to starvation, so adjust the following settings: inputBandwidthLimit= Your input limit outputBandwidthLimit= Your output limit Its in bytes/s. You should not allocate your whole bandwidth to freenet, leave at least some 10% to other traffic. You could also limit the threads used by adjusting the maximumThreads setting. - -- Roger Oksanen [EMAIL PROTECTED] +358 50 355 1990 CS Student at Helsinki UniversityPGP id 1B125A3E Homepage http://www.cs.helsinki.fi/u/raoksane/ -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.3 (GNU/Linux) iD8DBQFA+8CV78OZUBsSWj4RAmrtAJ9hcwJfIScktZbf224djNidhETolACgk/AY 2D594laB1rVZI7oOlwGm1ug= =5ffn -END PGP 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] Load
Roger Oksanen wrote: I run freenet niced at +10 on a 2x500MHz computer, load stays at 2-3 all the time. Ah yes, I forgot to mention that. It's niced at 19. Beats me how something that's niced 19 can bring the load to 5.00, but that's a different issue. I suspect the problem you have lies in the fact that freenet will eat ALL available bandwidth that you give it, which will lead to starvation, so adjust the following settings: inputBandwidthLimit= Your input limit outputBandwidthLimit= Your output limit That's done already, it's not where the problem lies. Both these settings are at 10240, calculated for a monthly consumption of about 50 GB. The machine has a 100 Mbit connection to the net, so starvation is out of the question. You could also limit the threads used by adjusting the maximumThreads setting. Reducing maximumThreads from default 120 to 60 had very little positive impact on the load. However, while I was there I noticed the overLoadlow parametre, which I had missed earlier. I set it to 0.8 but it dosn't work as advertised. After 35 minutes with this setting in effect, I'm looking at 9:57, 1 user, load average: 1.13, 1.74, 1.04 9:58, 1 user, load average: 1.53, 1.71, 1.07 10:00, 1 user, load average: 1.84, 1.77, 1.17 10:01, 1 user, load average: 3.04, 2.05, 1.31 10:05, 1 user, load average: 2.37, 2.39, 1.61 10:17, 1 user, load average: 5.49, 4.00, 2.69 10:26, 1 user, load average: 4.27, 4.20, 3.39 ./stop-freenet.sh 10:30, 1 user, load average: 0.15, 1.99, 2.66 If all averages are constantly above the overLoadlow limit and the one-minute average keeps increasing, then this setting is simply not being obeyed. Duh. I don't remember running a more aggressive piece of software, ever. A note to the developers: RAM is cheap. Working software is very expensive. Freedom is horrendously expensive. Sadly, this is an over-simplification and reality is more complex than that. The people who have money can buy freedom and don't need more RAM. The people who mostly need more RAM in order to have freedom are mainly those who can't afford the RAM. This is true on a national level, comparing the degree of repression and the financial situation of the average citizen in, say, China or Egypt to those in the US or Europe, and it is also true on the personal level; he who can pay a good team of lawyers will seldom need to fiddle with freenet. In my case, I rent a server somewhere for 39 euro per month. It's crappy hardware, but it's fully sufficient for all my needs and it's all I can afford anyway. To get better hardware where I have the bandwidth I'd have to double my expense. At home, where I have better hardware, I pay the traffic at the tune of 3 euro/GB. The sum of this equation is, unfortunately, one freenet node less. I do think that resource management would be a worthy priority for the project. Z -- Framtiden r som en babianrv, frggrann och full av skit. Arne Anka ___ 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] Load
On Mon, Jul 19, 2004 at 02:14:38PM +0200, Zenon Panoussis wrote: Hello everyone. I started a node on a machine with lots of bandwidth and a very lousy I/O subsystem. Not much else is going on on the machine, so without freenet the load is steadily between 0.01 and 0.10. When freenet runs, the load is constantly around 3.50, with peaks reaching well above 5.00. The system latency caused by these loads gets other stuff to malfunction; for instance, mail server queries to LDAP time out and result in service temporarily unavailable errors. Obviously, such things become a show stopper for freenet. Strange. What is your logLevel ? The machine is a Celeron 2.4 GHz with 512 MB RAM running RHEL3 with Sun java 1.4.2_05. I reduced maxNodeConnections to 85 from the default 200 in the hope to reduce the number of java instances, but that didn't help much. Won't make much difference. The setting you want is maximumThreads. I also set diagnosticsPath=/dev/null to stop the constant writing on disk (is there a better way to say disable stats, I don't need them?) and that didn't help much either. You do. The node uses them for estimating load. My long-lived unstable node's stats dir is 3MB, as is my long-lived stable node's stats dir. It's not a problem. So I'm asking for advice from those more experienced with freenet: what can I do to reduce the load? More specifically, how can I reduce the number of java instances running? With my current settings (most else at default), I have 96 java processes eating away everything on the machine and the machine itself too. Set maximumThreads=60 (remove the leading %), logLevel=error, and doCPULoad=true (is your node pegged on CPU, or just on I/O? what's the idle % typically?). -- 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] Load
On Mon, Jul 19, 2004 at 03:37:41PM +0300, Roger Oksanen wrote: I run freenet niced at +10 on a 2x500MHz computer, load stays at 2-3 all the time. I suspect the problem you have lies in the fact that freenet will eat ALL available bandwidth that you give it, which will lead to starvation, so adjust the following settings: inputBandwidthLimit= Your input limit outputBandwidthLimit= Your output limit Its in bytes/s. You should not allocate your whole bandwidth to freenet, leave at least some 10% to other traffic. I recommend allocating no more than half. Thus if you have an uplink of 256kbps, set: outputBytes=16000 I doubt that inputBandwidthLimit matters that much.. but if you have a symmetric connection, set it. You could also limit the threads used by adjusting the maximumThreads setting. -- 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] Load
On Mon, Jul 19, 2004 at 03:54:09PM +0200, Zenon Panoussis wrote: Roger Oksanen wrote: I run freenet niced at +10 on a 2x500MHz computer, load stays at 2-3 all the time. Ah yes, I forgot to mention that. It's niced at 19. Beats me how something that's niced 19 can bring the load to 5.00, but that's a different issue. I suspect the problem you have lies in the fact that freenet will eat ALL available bandwidth that you give it, which will lead to starvation, so adjust the following settings: inputBandwidthLimit= Your input limit outputBandwidthLimit= Your output limit That's done already, it's not where the problem lies. Both these settings are at 10240, calculated for a monthly consumption of about 50 GB. The machine has a 100 Mbit connection to the net, so starvation is out of the question. You could also limit the threads used by adjusting the maximumThreads setting. Reducing maximumThreads from default 120 to 60 had very little positive impact on the load. However, while I was there I noticed the overLoadlow parametre, which I had missed earlier. I set it to 0.8 but it dosn't work as advertised. After 35 minutes with this setting in effect, I'm looking at Please leave overloadLow alone. It messes up rate limiting. 9:57, 1 user, load average: 1.13, 1.74, 1.04 9:58, 1 user, load average: 1.53, 1.71, 1.07 10:00, 1 user, load average: 1.84, 1.77, 1.17 10:01, 1 user, load average: 3.04, 2.05, 1.31 10:05, 1 user, load average: 2.37, 2.39, 1.61 10:17, 1 user, load average: 5.49, 4.00, 2.69 10:26, 1 user, load average: 4.27, 4.20, 3.39 ./stop-freenet.sh 10:30, 1 user, load average: 0.15, 1.99, 2.66 If all averages are constantly above the overLoadlow limit and the one-minute average keeps increasing, then this setting is simply not being obeyed. Duh. I don't remember running a more aggressive piece of software, ever. A note to the developers: RAM is cheap. Working software is very expensive. Freedom is horrendously expensive. Sadly, this is an over-simplification and reality is more complex than that. The people who have money can buy freedom and don't need more RAM. The people who mostly need more RAM in order to have freedom are mainly those who can't afford the RAM. This is true on a national level, comparing the degree of repression and the financial situation of the average citizen in, say, China or Egypt to those in the US or Europe, and it is also true on the personal level; he who can pay a good team of lawyers will seldom need to fiddle with freenet. Yeah, whatever. IMHO we need to make freenet work before we make it work fast. And the timescale on making it work may be a timescale of years. In which case, RAM really is not an issue, for the time being. Also, we HAVE done significant work on reducing memory consumption in the last year or so. In my case, I rent a server somewhere for 39 euro per month. It's crappy hardware, but it's fully sufficient for all my needs and it's all I can afford anyway. To get better hardware where I have the bandwidth I'd have to double my expense. At home, where I have better hardware, I pay the traffic at the tune of 3 euro/GB. Yikes. Where do you live? Poland? Spain? Greece? Lithuania? :) The sum of this equation is, unfortunately, one freenet node less. I do think that resource management would be a worthy priority for the project. Along with the other 300 worthy priorities! -- 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] Load
Toad wrote: Strange. What is your logLevel ? Well, that's relative. The log level is set to debug, but the log file is a FIFO, where a simple perl script greps for URIs and dumps the rest. My idea was to feed those URIs to mnogosearch and create a non-anonymous search engine fo freenet. Won't make much difference. The setting you want is maximumThreads. I took them down from 120 to 60, saw hardly any difference at all. I also set diagnosticsPath=/dev/null to stop the constant writing on disk (is there a better way to say disable stats, I don't need them?) and that didn't help much either. You do. The node uses them for estimating load. My long-lived unstable node's stats dir is 3MB, as is my long-lived stable node's stats dir. It's not a problem. It's not the size of the stats on disk I want to avoid, but the extra I/O that comes from keeping them. Set maximumThreads=60 (remove the leading %), logLevel=error, and doCPULoad=true (is your node pegged on CPU, or just on I/O? what's the idle % typically?). OK, I'll give it one more try to see if it's the logging that does it. Of course, the logging is required for what I want to do, so I don't know how to get around the problem, if that's where it is. The normal load of the machine is around 0.10 an I/O is its big problem in general. Z -- Framtiden r som en babianrv, frggrann och full av skit. Arne Anka ___ 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] Load
On Mon, Jul 19, 2004 at 05:02:42PM +0200, Zenon Panoussis wrote: Toad wrote: Strange. What is your logLevel ? Well, that's relative. The log level is set to debug, but the log file is a FIFO, where a simple perl script greps for URIs and dumps the rest. My idea was to feed those URIs to mnogosearch and create a non-anonymous search engine fo freenet. Okay, this is your basic problem. I don't even use logLevel=debug any more. It produces crazy amounts of data, uses a lot of CPU, and the system cannot keep up. I doubt that feeding it through a FIFO will make that much difference - much of it is probably the generation side (as well as the syscalls). I recommend you set the following: logLevelDetail=freenet.client:debug That should catch all the URIs, probably. You don't need debug log level on everything else, and having it will really mess stuff up. Won't make much difference. The setting you want is maximumThreads. I took them down from 120 to 60, saw hardly any difference at all. I also set diagnosticsPath=/dev/null to stop the constant writing on disk (is there a better way to say disable stats, I don't need them?) and that didn't help much either. You do. The node uses them for estimating load. My long-lived unstable node's stats dir is 3MB, as is my long-lived stable node's stats dir. It's not a problem. It's not the size of the stats on disk I want to avoid, but the extra I/O that comes from keeping them. Which is minimal, if the OS has enough memory to batch the writes into one medium sized write every 30 seconds. Set maximumThreads=60 (remove the leading %), logLevel=error, and doCPULoad=true (is your node pegged on CPU, or just on I/O? what's the idle % typically?). OK, I'll give it one more try to see if it's the logging that does it. Of course, the logging is required for what I want to do, so I don't know how to get around the problem, if that's where it is. The normal load of the machine is around 0.10 an I/O is its big problem in general. I have seen similar problems when running logLevel=debug on my own nodes. The solution is to set logLevel=error and logLevelDetails for the individual subsystems you want to monitor. -- 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] Load
On Mon, Jul 19, 2004 at 09:03:47PM +0200, Zenon Panoussis wrote: Toad wrote: I recommend you set the following: logLevelDetail=freenet.client:debug You did uncomment it, right? Of course :) ...that now the URIs don't get logged. ' That's strange. What URIs were you after? Those of requests in transit and inserts. Because of the island-like nature of freenet publishing, traditional spidering won't get very far; you can't seed a search engine with a few sites and assume that you will find the entire network by following links. Monitoring requests and inserts musters the collective URI knowledge of one's peers and of their peers, so it could go a long way, especially if you can put together a mesh of URI-grabbing nodes in different places. In a word, NO. You cannot monitor the URIs of requests going through your node. If you could, you could decrypt the data. Then you could search it, sure. However you'd also be liable for it. It would probably make tracking authors down easier too. We have no way of knowing how big the other islands out there are... we can only spider out from known sites, and from publicly visible Frost traffic. The thing is, the lack of search capabilities reduces the useability of freenet Of course. There are ways to implement search, however. Sooner or later somebody will implement a good spider based anonymous search. This would probably have two components: 1. A spider, which would spider out from known freesites, scan NIMs, and Frost traffic, and insert index files. 2. A client, probably integrated into fproxy, which would fetch the index files that are appropriate to the search given. and, indirectly, compromises anonymity too. Not really. I can publish stuff anonymously all I want but, unless I post a URL somewhere, nobody is going to find my publications. Indeed. Thus we have NIMs, FreeMail and Frost within Freenet, and outside it we have Mixmaster remailers, IIP, I2P, various kinds of proxies and so on. Sadly some people use hushmail too, which is not exactly the safest option. But there are many possibilities. And conversely, if I'm looking for a piece of information that might well be on freenet, I won't find it without asking. No. You will search for it. Just like on the real Internet. You may use a search engine, or you may follow links. Either way, there are issues of trust which are ironically much more readily solved by hypertext than by just making everything searchable. A spammable search system is of little practical use. Especially for someone who's new to freenet and doesn't already have a set of bookmarks and starting points, the threshold for getting anywhere is pretty high despite the proxy bookmarks. How do you propose to protect against spam, and plain malicious content? A non-anonymous search engine on the web could solve part of both these problems and at the same time function as an invitation to freenet for non-freenet users. So spider! We're not stopping you. Freenet does not know the URIs of data that passes through the node, only those requested locally. It does know the requests that pass through the node. Nope. It doesn't. It only knows the routing keys, which are insufficient to decrypt the actual data. Any other URIs in the logs will be locally originated. Example: CHK@routing key,decrypt key/human readable key The node only knows the routing key. It does not know the decrypt key and therefore cannot decrypt the data. If it could, it could guess the human readable key. Last night, all of freenet for me was the few URIs that are published on freenetproject.org. This morning I had a whole long list in my logs, and through that I was able to start finding my way around. Would be interested to see some of this list. Are you running a public gateway? Are you fetching lots of stuff locally? That's how this idea of a search engine popped up and turned into a small project in itself. Most of these would go through freenet.client... some might go through freenet.node.states.FCP, and there are a few internal ones. I'll look. I'm grateful for any tips you might have. -- 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] Load
Toad wrote: The thing is, the lack of search capabilities reduces the useability of freenet Of course. There are ways to implement search, however. Sooner or later somebody will implement a good spider based anonymous search. I searched a bit on the web. At http://conferences.oreillynet.com/cs/p2pweb2001/view/e_sess/1669 I found someone claiming that searching freenet would be possible real soon, to quote: right about now. That was in 2001. At http://www.freenet.org.nz/search/ I found a totally defunct search engine, obviously based on the same principle I'm trying to apply now. I fully agree with you that anonymous search is much better than a non-anonymous. However, as I mentioned, the problem of anonymity has two sides: that of the publisher and that of the user. If a non-anonymous search solves one part without affecting the other, what's the harm of it? This would probably have two components: 1. A spider, which would spider out from known freesites, scan NIMs, and Frost traffic, and insert index files. 2. A client, probably integrated into fproxy, which would fetch the index files that are appropriate to the search given. You mean creating index files before a search has been made? Wouldn't that be highly inaccurate and/or produce massive volumes of indices? I can publish stuff anonymously all I want but, unless I post a URL somewhere, nobody is going to find my publications. Indeed. Thus we have NIMs, FreeMail and Frost within Freenet, and outside it we have Mixmaster remailers, IIP, I2P, various kinds of proxies and so on. Sadly some people use hushmail too, which is not exactly the safest option. But there are many possibilities. All this put together is still a *very* small world. If I'd find and publish, say, the Bush administration's plans to invade Cuba, or detailed information on Israel's chemical and biological weapons, I don't want this information to to reach the users of freenet and hushmail; I want it to reach the huge and clueless masses who watch CNN and use hotmail. And I also want to protect my anonymity damn well. The way to go? Publish on freenet and let automation, i.e. nobody, make the bridge to the web. How do you propose to protect against spam, and plain malicious content? I don't. I'm not Google. As you have already gathered, my financial capacity is enough to run a 39-euro server, but not a 78-euro one. Because of that, things get very simple: if I make a freenet search, it will be just as well or ill protected from spam and malicious content as freenet itself is. Freenet does not know the URIs of data that passes through the node, only those requested locally. It does know the requests that pass through the node. Nope. It doesn't. It only knows the routing keys, which are insufficient to decrypt the actual data. Any other URIs in the logs will be locally originated. Example: CHK@routing key,decrypt key/human readable key Uhm, there's something eluding me here. You know freenet's internals; I don't. If you say so, then so it is. Yet I stuck some of those URIs I found in my logs into my browser and got sites to which I had never been before. Taking what you say here for granted, the entire discussion up to this point is probably a meaningless exchange based on some misunderstanding on my part. But what? [URIs from logs] Would be interested to see some of this list. Duh. So am I by now, but with all the messing around today I deleted them. I can try again though. Are you running a public gateway? Are you fetching lots of stuff locally? Neither. Z -- Framtiden r som en babianrv, frggrann och full av skit. Arne Anka ___ 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] Load
I wrote: Taking what you say here for granted, the entire discussion up to this point is probably a meaningless exchange based on some misunderstanding on my part. But what? [URIs from logs] Would be interested to see some of this list. Duh. So am I by now, but with all the messing around today I deleted them. I can try again though. Now I know what the misunderstanding was. The working URIs I found in my logs come from the default bookmarks in the interface servlet. I had never visited them before, but they had passed my client anyway. Z -- Framtiden r som en babianrv, frggrann och full av skit. Arne Anka ___ 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] load average is too high
On Tuesday 06 April 2004 00:40, Conrad Sabatier wrote: I've been getting good results with the following: maxNodeConnections=128 maximumThreads=128 rtMaxNodes=256 targetMaxThreads=128 tfAbsoluteMaxThreads=128 How much RAM do you have? It said Reducing rtMaxNodes to 64. It still takes several (almost 7) minutes before it says starting node. phma -- li fi'u vu'u fi'u fi'u du li pa ___ 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] load average is too high
On Tuesday 06 April 2004 00:40, Conrad Sabatier wrote: I've been getting good results with the following: maxNodeConnections=128 maximumThreads=128 rtMaxNodes=256 targetMaxThreads=128 tfAbsoluteMaxThreads=128 The load average still shot up. It's 19.65 12 minutes after I started it, and the web page still hasn't come up. phma -- li fi'u vu'u fi'u fi'u du li pa ___ 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] load average is too high
On Tuesday 06 April 2004 06:44, Conrad J. Sabatier wrote: I have 512 MB of RAM. I'm curious now, how are you starting the node? What command are you using? ./start-freenet.sh I have 48 meg. What settings do you recommend? Could you try the following and let us know what you get: java -jar lib/freenet.jar --version Unable to access jarfile lib/freenet.jar There is no directory lib. If I type java -jar freenet.jar --version I get Fred version 0.5, protocol version STABLE-1.50 (build 5076, last good build 5074). Also, you may want to run 'java -jar lib/freenet.jar --config' to setup a new config file. It sounds like there may be a serious problem with your existing config. And at the risk of insulting your intelligence :-) ... don't forget that if you want to modify any default settings in your config file, you have to remove the leading % character. I told you they weren't set to anything. -- li fi'u vu'u fi'u fi'u du li pa ___ 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] load average is too high
On Tue, Apr 06, 2004 at 06:15:17AM -0400, Pierre Abbat wrote: On Tuesday 06 April 2004 00:40, Conrad Sabatier wrote: I've been getting good results with the following: maxNodeConnections=128 maximumThreads=128 rtMaxNodes=256 targetMaxThreads=128 tfAbsoluteMaxThreads=128 The load average still shot up. It's 19.65 12 minutes after I started it, and the web page still hasn't come up. What are the messageSendTime's like? I suppose you won't know if the web interface hasn't come up... What's the last thing in the logfile? phma -- 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] load average is too high
On Tue, Apr 06, 2004 at 06:15:17AM -0400, Pierre Abbat wrote: On Tuesday 06 April 2004 00:40, Conrad Sabatier wrote: I've been getting good results with the following: maxNodeConnections=128 maximumThreads=128 rtMaxNodes=256 targetMaxThreads=128 tfAbsoluteMaxThreads=128 You might want to try logLevel=error also. And reduce the thread settings, since your machine is slow - say maximumThreads=64, rtMaxNodes=32, targetMaxThreads=128, tfAbsoluteMaxThreads=192 The load average still shot up. It's 19.65 12 minutes after I started it, and the web page still hasn't come up. phma -- li fi'u vu'u fi'u fi'u du li pa ___ 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] load average is too high
You could try doCPULoad=true, but you'd have to turn off the background CPU hog. That makes the node tell other nodes to send it fewer queries until its CPU usage is reasonable. What is the messageSendTime? On the General page? On Tue, Apr 06, 2004 at 02:14:48PM -0400, Pierre Abbat wrote: On Tuesday 06 April 2004 12:28, Toad wrote: What are the messageSendTime's like? I suppose you won't know if the web interface hasn't come up... What's the last thing in the logfile? The web interface comes up but is very slow. I turned the node off, as it was just hiking the load average and not doing much useful. Here's the last line in the logfile: Apr 6, 2004 7:24:49 AM (freenet.node.states.request.TransferReply, YThread-85, NORMAL): Failed to send data with CB 0x83 (CB_RECV_CONN_DIED), for [EMAIL PROTECTED]: key=accec1d9a0abcc389ed67f127225edc0c4a4201b0f0203, hopsToLive=8, id=a4f3f7124f06401e, [EMAIL PROTECTED] (accec1d9a0abcc389ed67f127225edc0c4a4201b0f0203,request), ft=freenet.client.InternalClient$InternalGetToken:[EMAIL PROTECTED],key=freenet:[EMAIL PROTECTED]/4071f300-FreenetHelp,skipBytes=0, orig=null, last=Peer [DSA(c3cf b843 3201 e6b6 0989 8124 4d3f b35a 9615 c39c) @ 195.56.102.112:6127 (1/3)], routedTime=1081250326751, replyTime=1081250335817, outwardSender=null I might could put the node on my laptop, which has 256 meg, but I sometimes take the laptop to LUG meetings, to work, or elsewhere. phma -- li fi'u vu'u fi'u fi'u du li pa ___ 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] load average is too high
On Tuesday 06 April 2004 14:17, Toad wrote: You could try doCPULoad=true, but you'd have to turn off the background CPU hog. That makes the node tell other nodes to send it fewer queries until its CPU usage is reasonable. What is the messageSendTime? On the General page? It took several minutes for the General page to come up. messageSendTimeRequest is 0, which probably doesn't tell you anything, so here's the page. phma -- li fi'u vu'u fi'u fi'u du li paTitle: General Information Freenet General Information [Switchtoadvancedmode][Savecurrentmode] Node Information Web Interface Bookmark Manager Performance GeneralInformation Networking CurrentDownloads SpreadFreenet Documentation READMEfile Node Information From here you can view information about what is going on inside your node. Select from the options in the menu to the left. Version Information Node Version0.5Protocol VersionSTABLE-1.50Build Number5076CVS Revision1.90.2.50.2.103 Uptime 29 minutes Load Current routingTime0msCurrent messageSendTimeRequest0msPooled threads running jobs71 (55.5%)Pooled threads which are idle36Current upstream bandwidth usage183 bytes/second (1.5%)Current estimated load for QueryReject purposes55%Current estimated load for rate limiting55.5%Reason for load:Load due to thread limit = 55.5%Load due to routingTime = 10% = 10,000% / 100,000% <= overloadLow (50%)Load due to messageSendTimeRequest = 10% = 10,000% / 100,000% <= overloadLow (50%)Load due to output bandwidth limiting = 1.5% because outputBytes(11003) <= limit (589824.009 ) = outLimitCutoff (0.8) * outputBandwidthLimit (12288) * 60Load due to expected inbound transfers: 0% because: 0.0 req/hr * 1.5333992799054395E-4 (pTransfer) * 35.0 bytes = 0 bytes/hr expected from current requests, but maxInputBytes/minute = 1966080 * 60 * 1.1 = 129761280 bytes/hr targetEstimated external pSearchFailed (based only on QueryRejections due to load):0.0Current estimated requests per hour (based on last 10 mins):0.0Current global quota (requests per hour):0.0Highest seen bytes downloaded in one minute:11986Current outgoing request rate0.0 ___ 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] load average is too high
On Tue, Apr 06, 2004 at 03:23:14PM -0400, Pierre Abbat wrote: On Tuesday 06 April 2004 14:17, Toad wrote: You could try doCPULoad=true, but you'd have to turn off the background CPU hog. That makes the node tell other nodes to send it fewer queries until its CPU usage is reasonable. What is the messageSendTime? On the General page? It took several minutes for the General page to come up. messageSendTimeRequest is 0, which probably doesn't tell you anything, so here's the page. Hmm. It's really struggling, even though it's not doing anything... I dunno what we can do about it... phma -- 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] load average is too high
On Tuesday 06 April 2004 15:30, Toad wrote: Hmm. It's really struggling, even though it's not doing anything... I dunno what we can do about it... I moved it to the laptop, and it seems to be doing fairly well. Oddly, there is little activity on my Freenet port, as seen in tcpdump, even when I'm browsing it. phma -- li fi'u vu'u fi'u fi'u du li pa ___ 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] load average is too high
It took several minutes for the General page to come up. messageSendTimeRequest is 0, which probably doesn't tell you anything, so here's the page. Hmm. It's really struggling, even though it's not doing anything... I dunno what we can do about it... Produce a couple of full stackdumps to see where the threads are stuck? /N ___ 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] load average is too high
I'm running Freenet on a 48MB machine which is also running mprime. The load average varies from 7 to 14, whereas before I started Freenet the load average was around 1.3. I'm using j2re 1.5.0-beta, which I downloaded today from Sun. top shows dozens of java processes, all the same size and all varying together. Is there a parameter in freenet.conf that I can tweak to get the load average down? phma -- li fi'u vu'u fi'u fi'u du li pa ___ 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] Load
Persistent node (525) here showing 100% load (106% and change right now) most of the time. Is this something to worry about? And if it is, is there any information you'd like to see? -- Democracies die behind closed doors. - Judge Damon Keith ___ support mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support
Re: [freenet-support] Load
mine has been up to 89% at times. i think this is due to the /. effect at slashdot.org once all the influx of new users .. uuhh.. like me.. it should go back down. if the 100 % stays too long, just restart the freenet :P - Original Message - From: Doug Bostrom [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Monday, October 28, 2002 3:38 PM Subject: [freenet-support] Load Persistent node (525) here showing 100% load (106% and change right now) most of the time. Is this something to worry about? And if it is, is there any information you'd like to see? -- Democracies die behind closed doors. - Judge Damon Keith ___ support mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support ___ support mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support
Re: [freenet-support] Load
On Mon, Oct 28, 2002 at 06:38:25PM -0500, Doug Bostrom wrote: Persistent node (525) here showing 100% load (106% and change right now) most of the time. Is this something to worry about? And if it is, is there any information you'd like to see? Yes, but we are aware of the problem. -- 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 11/11/02. http://freenetproject.org/ msg01685/pgp0.pgp Description: PGP signature
