On Wednesday 28 January 2009 07:24, 3BUIb3S50i 3BUIb3S50i wrote: > Don s...@tgzmfuebneqarnhpyqj4of3-s21b0utliyfallq0bw8 wrote: > > One of the recent Freenet builds caused my Freenet node end up with > > almost constant 100% CPU usage. This trouble began somewhere inbetween > > 1199-1203 releases. My node used to work very nicely before that. Is > > anyone else experiencing this problem? > > > > I can post info Stats page and Freenet logs to help in diagnosing the > > problem if needed. Just let me know which parts in particular you're > > interested in. Regarding wrapper.log, I found nothing of use in there, > > just the node (re)start/stop related activity.
Please determine whether this is a memory-related problem. Add the following to your wrapper.conf, and then view the file freenet.loggc: wrapper.java.additional.3=-Xloggc:freenet.loggc If it is showing Full GC's very frequently (once a second or more), then there is a memory problem. If not, the problem is something else... A stack dump might also be interesting. If you are doing a big download or upload, there will inevitably be periods when Freenet uses lots of CPU for FEC encoding/decoding, or for compression/decompression; but this is generally run at a low priority. > > > > I even tried complete reinstall of the node with new datastore, but to no > > avail. Datastore type is salt-hash. Is the download queue empty? > > > > My specs: > > JVM Info > > > > * Used Java memory: 159 MiB > > * Allocated Java memory: 244 MiB > > * Maximum Java memory: 508 MiB > > * Running threads: 280/500 May not be a memory problem then ... Please do the above steps anyway. If it isn't a memory problem, a stack dump might be useful (note that it may have sensitive information in it sometimes...) > > * Available CPUs: 1 > > * Java Version: 1.6.0_10 > > * JVM Vendor: Sun Microsystems Inc. > > * JVM Version: 11.0-b15 > > * OS Name: Linux > > * OS Version: 2.6.27-9-generic > > * OS Architecture: i386 > > bu...@uv~lazhgpqpilgjutc2bd~mmdm5r1pjkyss3kznuixm wrote: > > tommy...@eefdujdzxdwxl0qusx0cjofgmjbvd3df4ty61pzy8y8 wrote: > > > >> myident...@1qowk8lzyeynusi0ygamwcd6ox80xqkkr8kcs6pmj5q schrieb: > >>> Has happened a couple times in the last couple days. All through put simply > >>> stops with my node. Once I simply waited a long time and it started to work > >>> again but the speed was very, very slow so I restarted the node. > >>> > >>> One time confirmed for sure it died for 6 hours till I got back to my computer > >>> and restarted the node. > >>> > >>> After restarting the node it seems to work fine again till it dies. This is not > >>> common maybe it has happened 3-4 times since the last update. > >>> > >>> I run a fast computer with lots of memory. 1 gig given to freenet. > >>> > >>> I will continue to monitor this and report back if it continues. > >> > >> Would be nice to have some data out of wrapper.log or logs/*, else it is hard to find the problem > >> and fix it. Also other things like node stats can be interesting. > > > > same here, i think it's a memory leak That is easily confirmed, see the procedure mentioned above.
pgpVSRXZA1hh0.pgp
Description: PGP signature
_______________________________________________ Support mailing list [email protected] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[email protected]?subject=unsubscribe
