On Fri, Jul 02, 2004 at 10:06:00PM -0400, Nicholas Sturm wrote: > Toad wrote: > > >On Fri, Jul 02, 2004 at 10:56:13AM -0400, Paul wrote: > > > > > >>Many dialup connections are regularly reset. They probably would have > >>locked his account if he had gone over bandwidth or connection time. > >>Getting disconnected is just a fact of life. > >> > >> > > > >Okay so it's not caused by Freenet? Good. > > > > > >>~Paul > >> > >>On Thu, 1 Jul 2004 16:08:53 +0100, Toad <[EMAIL PROTECTED]> wrote: > >> > >> > >>>You have no idea WHY it lost the connection to the ISP? Did they contact > >>>you to complain about bandwidth usage or anything? How do you connect to > >>>the internet? Has that changed recently? > >>> > >>> > >>>On Wed, Jun 30, 2004 at 06:38:37PM -0400, Nicholas Sturm wrote: > >>> > >>> > >>>>Restarted freenet last night. Slow to make contacts but by an hour > >>>>later > >>>>62 were open and freenet seemed to be behaving nicely together with SETI > >>>>running. CPU at 100% but behavior was what I now call "normal." > >>>> > >>>>At about noon today I checked and data transferred was many megs and > >>>>messages were at about 15,000 on two most active connections. Checked > >>>>mail > >>>>and came back about an hour later to find that SETI had transmitted > >>>>results > >>>>and was "not" maximized any longer. Then I realized that system seemed > >>>>inactive. ISP had closed connection and I had some difficulty getting > >>>>anything to respond. Finally after right click on rabbit in tray, I was > >>>>able to open popup window and stopped freenet. Few things started to > >>>>show > >>>>apparent activity and I could then maximize SETI and it had completed > >>>>about > >>>>6% of a job after sending and bringing down a new job. > >>>> > >>>>I then reconnected and restarted freenet and function seemed to return. > >>>>I > >>>>checked the log and it showed a very long segment of failures. (Log > >>>>was at > >>>>about 2.5 megs.) Errors continued abundantly as I expected since > >>>>contact > >>>>had been lost from other nodes for some time. Shut down. (To do other > >>>>work.) About three hours later I tried to restart, but experience > >>>>little > >>>>success. Log hung when I tried to work back through the log (problem > >>>>here?). Finally shut down OS and restarted. > >>>> > >>>>The apparent hang from shut down of ISP connection I had not observed > >>>>before (not to say that I actually know it never did). Is this common? > >>>>Will freenet do this when only it is producing high CPU usage or could > >>>>it > >>>>be because two programs were trying to work at maximum level (SETI & > >>>>freenet)? Should freenet not detect loss of internet connection and > >>>>not go > >>>>blindly on with unsuccessful high usage? I would think that when > >>>>finally > >>>>operating this should not be allowed to happen as ISP closure would > >>>>certainly not be uncommon with large numbers of nodes running (even if > >>>>it > >>>>only happens with multiple high demands on CPU). > >>>> > >>>>Recently checked thread usage and seems seldom to go beyond 700 even > >>>>when > >>>>system very busy. > >>>> > >>>>Oh, 2KWin and Sun Java (recent). 256 memory. Dial up connection. > >>>>~18.6 > >>>>hard drives capacity each of two(C: pretty high, about 1.5 gig open, D: > >>>>with about 4-5 gig open). What else important?
> Apart from describing the behavior, the point I was trying to make was > that if this happens when the system is in general operation, many may > be affected by freezing of the system -- particularly if ISPs become > even more accustomed to doing it -- because those who have a frozen > system every morning will likely be inclined not to support nodes except > perhaps in the transient state. And then integration into the system > for access to information will be unlikely to be very popular. IF the > node freed itself when there is no active Internet link, i.e., when into > a paused state to avoid freezing, it might remain more useful as a > communication medium. Even better perhaps would be a timed pause with a > reconnection after a moderate > pause period -- 1 minute, 5 minutes, 10 minutes, or 30 minutes, say. > ????? It does seem like a potential problem. Well, I don't see why the node would cause a general hang. And in fact, I'd expect that the node's high CPU usage would go away quite quickly in the absence of an internet connection, given all the backoff... evidently this doesn't happen. Any idea why? What does it show on the Environment page when in this state? -- 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]
