I have found on recent 0.5 pre-releases (including pre2) that after a while (running overnight) the load of the node is pegged at 100% while browsing and my system is slowed down dramatically. On running lsof -i on my linux system it looks like there are a few separate instances of Freenet running on my system. Symptoms of this are many repetitions of a java process listening on 8888 and 8481 and my node port 21872 (111 repetitions of each LISTEN entry to be exact). I can assure you that I did not start up my Freenet node 111 times. I was under the impression that there is one LISTENing process per FCP/FNP port that spawns other connections on other ports as necessary.
Environment after closing down my Freenet browser (an possibly freeing up threads): JVM Vendor Sun Microsystems Inc. <http://java.sun.com/> JVM Name Java HotSpot(TM) Server VM JVM Version 1.3.1_02-b02 Operating System Linux OS Version 2.4.18-10 Total size of the data store 200 MB Free space in the data store 2,416 KB Used space in the data store 202,384 KB Total pooled threads 111 Available pooled threads 25 Pooled threads in use 86 Is this how things are supposed to be? I notice that the Total pooled threads matches the 111 Listening ports and that could have been the number in use at the time that I did the lsof -i . I was under the impression that the number of java threads would be equal to the number of java processes accessing ports or the number of java process entries in the ps list. As it is now, the lsof and ps commands list many many many times more then the pooled thread number including many repeats (111 entries for each unique connection to other nodes). Oh ... another thing that would prevent kibibyte vs. kilobyte questions would be to put "Total size of the data store" in the same units as the "Free space ..." and "Used space ...". That is more of a cosmetic thing though. Mike _______________________________________________ devl mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
