Re: [freenet-dev] stable net kicking ass
Take this suggestion for what its worth (as my opinion currently holds none in this conversation.) Don't make any release with the current stable build, nor any NGR build for that matter. In my humble opinion this would be the single biggest mistake the project could possibly make. Ian, You often make reference to the experience a first time user has with freenet. I concur and would hope others reading this thread will understand how valuable their first impression of freenet can be to the project. The reality is the current stable build is, *cough* most likely on par with ngr's performance and reliability in delivering content to the user. I've used both quite a bit and for the life of me I cant find any difference between the two. There have been a number of people reporting success with the current stable network, unfortunately this wonderful sensation likely only lasts for about an hour. I'll make mention that the original author of this thread spent some time tonight on IRC describing the pains he was having trying to relieve the monumental load placed on his node with no success as far as I could tell. If any person on this list does not believe me.. I would suggest getting yourself on stable and letting your node run for a few hours (please dont get on here and tell me how great your node is running after 15 minutes and 300 queries) The proof is in the pudding and unfortunately I have the pudding sitting right in front of me. I would be more than happy to be proven wrong. I'm dreaming that there are no load problems? Content retrieval is next to impossible? Wish I were wrong about those two but I'm not. My suggestion. Rebrand one of the golden age builds (ie. pre-pcache/nio) as the New and Improved Freenet... release that to the public, make $$$, and continue work on ngr. Ignore all of this if there's some master to plan to fix all of the load and retrieval problems before the imminent release... Dont let this mistake happen for the sake of the projects future. Shalom, Robert. On Sun, 2003-11-30 at 19:18, Ian Clarke wrote: Zlatin Balevsky wrote: I still think its a good idea to invest some toadtime into polishing it before 0.5.3, but so far its kicking ass. That is good news, and yes - I suspect there are a few things that need to be tidied up (for example, didn't someone mention that acquisition of new refs was borked?). Ideal scenario is to get 0.5.3 working nicely (perhaps releasing 0.5.3rc1 in a week), release it, get $$$, and then start figuring out what is wrong with NGR using a careful scientific process rather than the somewhat haphazard approach we have been taking. Ian. ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] stable net kicking ass
On Tue, Dec 02, 2003 at 03:30:39AM -0500, Robert Soros wrote: Take this suggestion for what its worth (as my opinion currently holds none in this conversation.) Don't make any release with the current stable build, nor any NGR build for that matter. In my humble opinion this would be the single biggest mistake the project could possibly make. Ian, You often make reference to the experience a first time user has with freenet. I concur and would hope others reading this thread will understand how valuable their first impression of freenet can be to the project. The reality is the current stable build is, *cough* most likely on par with ngr's performance and reliability in delivering content to the user. I've used both quite a bit and for the life of me I cant find any difference between the two. There have been a number of people reporting success with the current stable network, unfortunately this wonderful sensation likely only lasts for about an hour. I'll make mention that the original author of this thread spent some time tonight on IRC describing the pains he was having trying to relieve the monumental load placed on his node with no success as far as I could tell. If any person on this list does not believe me.. I would suggest getting yourself on stable and letting your node run for a few hours (please dont get on here and tell me how great your node is running after 15 minutes and 300 queries) The proof is in the pudding and unfortunately I have the pudding sitting right in front of me. I would be more than happy to be proven wrong. I'm dreaming that there are no load problems? Content retrieval is next to impossible? Wish I were wrong about those two but I'm not. My suggestion. Rebrand one of the golden age builds (ie. pre-pcache/nio) as the New and Improved Freenet... release that to the public, make $$$, and continue work on ngr. Ignore all of this if there's some master to plan to fix all of the load and retrieval problems before the imminent release... Dont let this mistake happen for the sake of the projects future. There are tons of ideas for fixing them. It's a matter of whether any of them work. I have heard fairly mixed reports on the new stable - scum says it's crap especially for inserts, but routingSuccessRatio is higher. For example. Shalom, Robert. On Sun, 2003-11-30 at 19:18, Ian Clarke wrote: Zlatin Balevsky wrote: I still think its a good idea to invest some toadtime into polishing it That is good news, and yes - I suspect there are a few things that need to be tidied up (for example, didn't someone mention that acquisition of new refs was borked?). Ideal scenario is to get 0.5.3 working nicely (perhaps releasing 0.5.3rc1 in a week), release it, get $$$, and then start figuring out what is wrong with NGR using a careful scientific process rather than the somewhat haphazard approach we have been taking. Ian. -- 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 ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
RE: [freenet-dev] stable net kicking ass
Quoting Edward J. Huff [EMAIL PROTECTED]: I asked how these connections got established. It seems we _never_ send requests to nodes not in the RT. Ergo, they were in the RT but got dropped. I haven't checked the code to make certain, however. I didn't see any questions asked about these connections, but if you can point me in the direction to work out how or why they are established I'll keep an eye on my node to see whats going on with it. Pete ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] stable net kicking ass
Connections are established in two situations. 1. When we have the node in our rt we open a connection to it since we might want to route requests to it. 2. When we need to send a message to that node.. (A sends request to B over connection AB, connection AB is closed for one reason or another, B needs to send the answer to A, connection BA is established) /N - Original Message - From: Peter Soden [EMAIL PROTECTED] To: Discussion of development issues [EMAIL PROTECTED] Sent: Monday, December 01, 2003 3:12 PM Subject: RE: [freenet-dev] stable net kicking ass Quoting Edward J. Huff [EMAIL PROTECTED]: I asked how these connections got established. It seems we _never_ send requests to nodes not in the RT. Ergo, they were in the RT but got dropped. I haven't checked the code to make certain, however. I didn't see any questions asked about these connections, but if you can point me in the direction to work out how or why they are established I'll keep an eye on my node to see whats going on with it. Pete ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] stable net kicking ass
The code is designed to do this, if lots of RT nodes are backed off in NGR... it wasn't implemented the same way in CPAlgoRT routing. Oh, and some of them will be transients. On Mon, Dec 01, 2003 at 06:31:20AM -, Pete wrote: Just looked at my nodes stats before I close it down for the time being and I noticed this figure Outbound connections that are to peers not in the routingtable 20.37037% Why aren't these peers being added to the routing table? They obviously are capable of dealing with request which Thses nodes Node references requesting ARKs 46 Obviously aren't, I know churning the rt is not exactly a great thing to do, but I think it may help in the long run if we drop nodes that really are behaving badly for ones that are working, just a suggestion... Pete -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Ian Clarke Sent: 01 December 2003 00:19 To: Discussion of development issues Subject: Re: [freenet-dev] stable net kicking ass Zlatin Balevsky wrote: I still think its a good idea to invest some toadtime into polishing it before 0.5.3, but so far its kicking ass. That is good news, and yes - I suspect there are a few things that need to be tidied up (for example, didn't someone mention that acquisition of new refs was borked?). Ideal scenario is to get 0.5.3 working nicely (perhaps releasing 0.5.3rc1 in a week), release it, get $$$, and then start figuring out what is wrong with NGR using a careful scientific process rather than the somewhat haphazard approach we have been taking. Ian. ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl -- 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 ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] stable net kicking ass
On Mon, Dec 01, 2003 at 08:46:04AM -0500, Edward J. Huff wrote: On Mon, 2003-12-01 at 01:31, Pete wrote: Just looked at my nodes stats before I close it down for the time being and I noticed this figure Outbound connections that are to peers not in the routingtable 20.37037% Why aren't these peers being added to the routing table? They obviously are capable of dealing with request which [these nodes obviously] aren't, I know churning the rt is not exactly a great thing to do, but I think it may help in the long run if we drop nodes that really are behaving badly for ones that are working, just a suggestion... I asked how these connections got established. It seems we _never_ send requests to nodes not in the RT. Ergo, they were in the RT but got dropped. I haven't checked the code to make certain, however. No. They sent US requests. -- Ed Huff -- 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 ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
[freenet-dev] stable net kicking ass
Ok, yesterday it didn't work - nodes didn't learn about each other, it was busy, etc. Today however, the request success ratio was more than 10% for a brand new node (18 out of 124 externally requested keys successful), and frost is chugging along (recorded a download speed of 100k at some point). And an insert of a small file with htl 21 took just few seconds (hope nobody is running a black hole). And nodes learn about each other all of a sudden... I still think its a good idea to invest some toadtime into polishing it before 0.5.3, but so far its kicking ass. pgp0.pgp Description: PGP signature ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] stable net kicking ass
I still think its a good idea to invest some toadtime into polishing it before 0.5.3, but so far its kicking ass. Yeah, just downloaded almost 300MB in half a day! That's a new record for me with Freenet. And we must consider that my stable node is completely new, I wiped out everything I had before (but the datastore). However, even if my node seems to get to other node very good, others are not going to me as well: in the same time, I've got only 53 inbound request search keys, a little too low for 5 hours of work. Aureliano Rama Corso di Laurea in Informatica, Pisa ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] stable net kicking ass
Zlatin Balevsky wrote: I still think its a good idea to invest some toadtime into polishing it before 0.5.3, but so far its kicking ass. That is good news, and yes - I suspect there are a few things that need to be tidied up (for example, didn't someone mention that acquisition of new refs was borked?). Ideal scenario is to get 0.5.3 working nicely (perhaps releasing 0.5.3rc1 in a week), release it, get $$$, and then start figuring out what is wrong with NGR using a careful scientific process rather than the somewhat haphazard approach we have been taking. Ian. ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
RE: [freenet-dev] stable net kicking ass
I'm not sure the acquisition of new ref's is borked, I've watched my rt shrink and regrow, so I'm pretty confident that that's not an issue, but I have noticed that every ref in my rt eventually backs off, which isn't very helpful to the cause, it also seems nodes aren't passing data between each other freely watching the convo is #freenet, (I'm sure zab will comment on this later). I think someone mentioned that some of the ref harvesters are down that could be a problem, as everyone installing/upgrading will be starting with pretty much the same refs and punishing the nodes in them. There are also far to many nodes which have arks that never seem to be found, (49 currently in my rt) that can't be helping the cause as they don't seem to ever get dropped :( Anyway nice to see some positive steps being taken in the development of freenet again Pete -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Ian Clarke Sent: 01 December 2003 00:19 To: Discussion of development issues Subject: Re: [freenet-dev] stable net kicking ass Zlatin Balevsky wrote: I still think its a good idea to invest some toadtime into polishing it before 0.5.3, but so far its kicking ass. That is good news, and yes - I suspect there are a few things that need to be tidied up (for example, didn't someone mention that acquisition of new refs was borked?). Ideal scenario is to get 0.5.3 working nicely (perhaps releasing 0.5.3rc1 in a week), release it, get $$$, and then start figuring out what is wrong with NGR using a careful scientific process rather than the somewhat haphazard approach we have been taking. Ian. ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] stable net kicking ass
I have always thought that the whole NGR thing was misguided. Freenet works because of keyspace specialization. NGR trashes keyspace specialization in favor of speed, which reduces Freenet from an medium speed expressway to a grid of city-streets with traffic-lights and congestion. You might be able to get across town faster by taking the side streets -- until everyone else starts taking the side streets also and your route becomes a parking-lot from the cross-traffic. IMHO, Toad's time would be better spent (post 0.5.3 of course) on MUXing and other minor optimizations that will benefit the project more for 0.6. In the meantime, other dev's can work on hammering out the bugs in NGR (0.7 anyone? :-p) - Original Message - From: Ian Clarke [EMAIL PROTECTED] To: Discussion of development issues [EMAIL PROTECTED] Sent: Sunday, November 30, 2003 7:18 PM Subject: Re: [freenet-dev] stable net kicking ass Zlatin Balevsky wrote: I still think its a good idea to invest some toadtime into polishing it before 0.5.3, but so far its kicking ass. That is good news, and yes - I suspect there are a few things that need to be tidied up (for example, didn't someone mention that acquisition of new refs was borked?). Ideal scenario is to get 0.5.3 working nicely (perhaps releasing 0.5.3rc1 in a week), release it, get $$$, and then start figuring out what is wrong with NGR using a careful scientific process rather than the somewhat haphazard approach we have been taking. Ian. ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
RE: [freenet-dev] stable net kicking ass
Just looked at my nodes stats before I close it down for the time being and I noticed this figure Outbound connections that are to peers not in the routingtable 20.37037% Why aren't these peers being added to the routing table? They obviously are capable of dealing with request which Thses nodes Node references requesting ARKs 46 Obviously aren't, I know churning the rt is not exactly a great thing to do, but I think it may help in the long run if we drop nodes that really are behaving badly for ones that are working, just a suggestion... Pete -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Ian Clarke Sent: 01 December 2003 00:19 To: Discussion of development issues Subject: Re: [freenet-dev] stable net kicking ass Zlatin Balevsky wrote: I still think its a good idea to invest some toadtime into polishing it before 0.5.3, but so far its kicking ass. That is good news, and yes - I suspect there are a few things that need to be tidied up (for example, didn't someone mention that acquisition of new refs was borked?). Ideal scenario is to get 0.5.3 working nicely (perhaps releasing 0.5.3rc1 in a week), release it, get $$$, and then start figuring out what is wrong with NGR using a careful scientific process rather than the somewhat haphazard approach we have been taking. Ian. ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.543 / Virus Database: 337 - Release Date: 21/11/2003 ___ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl