On Monday 05 January 2009 04:53, Dennis Nezic wrote: > On Fri, 2 Jan 2009 12:11:07 -0500, Dennis Nezic > <dennisn at dennisn.dyndns.org> wrote: > > > On Wed, 31 Dec 2008 14:41:42 +0000, Matthew Toseland > > <toad at amphibian.dyndns.org> wrote: > > > > > On Friday 26 December 2008 17:50, Dennis Nezic wrote: > > > > After a day or so, with no problems mentioned in wrapper.log, my > > > > wrapper will suddenly crash: > > > > > > > > "JVM appears hung: Timed out waiting for signal from JVM." > > > > "JVM did not exit on request, terminated" > > > > "JVM received a signal SIGKILL (9)." > > > > "Reloading Wrapper configuration..." > > > > > > > > It's not a new problem. > > > > > > > > Does this happen to anyone else? > > > > > > > > Here is a 24h graph of the free memory on my computer, during my > > > > latest jvm crash. I have other things running, but most of the > > > > activity is probably due to freenet. The crash occurred around > > > > 7am, and you can see how just over 100M is freed up... though, > > > > this is barely half of what I alot to it (220M). So I don't think > > > > it's an out-of-memory problem. > > > > > > > > http://dennisn.dyndns.org/guest/pubstuff/freenetcrashes-freememory.png > > > > > > > > But what else could crash the jvm? Maybe it's a bug in the > > > > wrapper? System CPU activity is acceptably low and normal (~25% > > > > average usage)--I can't imagine how it could hang the jvm for a > > > > few minutes. > > > > > > You can verify whether it is in fact a memory problem by adding to > > > your wrapper.log : > > > > > > wrapper.java.additional.3=-Xloggc:freenet.loggc > > > > > > Then tail -f freenet.loggc : when it crashes, do you see lots of > > > Full GC's, very frequently (approx every second)? > > > > I have a growing suspicion that there's a bug in either the wrapper, > > or freenet's interface with the wrapper. I am currently running > > without any wrapper, and I already have a few days of > > uptime--normally it would crash after about a day. Fingers crossed. > > Nevermind :\. After almost a week of uptime, it crashed, or, more > likely, stopped itself. Here's what wrapper.log said: > > Restarting node: MessageCore froze for 3 minutes! > Exiting on deadlock.... > Restarting node: MessageCore froze for 3 minutes! > Exiting on deadlock.... > Restarting node: PacketSender froze for 3 minutes! (264273) > Exiting on deadlock.... > Goodbye. from freenet.node.Node (USM deadlock) > Goodbye. from freenet.node.Node (PacketSender deadlock) > run() exiting for UdpSocketHandler..... > Goodbye. from freenet.node.Node (USM deadlock) > > It then goes on to successfully close the key stores. > > Well, at least it doesn't appear to be a memory problem? It was > steadily hovering at the same level (below the max i set for it) for > days. Then just suddently it shutdown.
Did you examine the freenet.loggc? Was it doing lots of Full GC's? I would be very surprised if this was not a memory problem based on past experience! > > Ideas? -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 827 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/support/attachments/20090105/0b4a607b/attachment.pgp>
