[freenet-support] Freenet 0.7 build 1169
Freenet 0.7 build 1169 is now available. Please update. No massive changes, but some important bugfixes which hopefully will fix some of the connectivity problems that have been reported recently: - Fix announcing to the same node more than once, disconnecting, wasting lots of announcements. (Thanks sdiz) - Don't do DNS lookups constantly on the packet sender thread, this caused havoc with a node with a nonexistent DNS address. - Logging to try to track down problems with updating. - Make announcements slightly more tolerant of lag. - Make the salted-hash datastore the default for new nodes (but not for old nodes). The salted-hash datastore, written by sdiz, is a massive improvement in most respects on the old bdbje datastore. It features random replacement (which has shown in simulations to slightly improve network performance), uses much less CPU, is much more robust, and uses less in-JVM memory. However, by default it uses a Bloom filter to minimise disk I/O and improve performance; this uses 1/2048th of the size of the store, memory mapped to a disk file; it doesn't count towards the JVM memory limit. You can turn this off if seeking is not a problem (e.g. on a flash disk). Note that the salted-hash is only default for nodes that don't already have a datastore type: old nodes will not be automatically converted, since conversion takes a huge amount of disk space. This is another build with minimal changes - trunk has a number of major changes, but is not yet ready to release, so this build is based on 1166/1167/1168 rather than on the current trunk code. Please report any bugs you find! Thanks. -- 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/20081106/fb1a8325/attachment.pgp>
[freenet-support] Node won't connect to peers 1167/68
Okay, the PacketSender resent-and-not-acked problem is probably caused by a bug fixed in 1169. Unfortunately we will have to wait for most of the network to update before this will make a big impact. On Thursday 06 November 2008 02:43, Dennis Nezic wrote: > > On Wed, 5 Nov 2008 22:00:52 +, > > Matthew Toseland wrote: > > > > It could very well be. Lack of memory resulting in constant garbage > > collection could easily cause a timeout. > > I really don't think it's low memory. I increased > wrapper.java.maxmemory to 300, and even though only about 180MB is > being used (only 250MB is allocated.. ie. 50MB hasn't been touched > yet), it still crashes. (Also, varying maxmemory between 180 and 300 > doesn't really affect the interval between crashes.) > > Just to further stress the severity of the PacketSender "packet > not acked" errors that I'm getting, my freenet-previous.log is 80MB! > (0.1MB of which is another RequestSender "Waited too long for a > blocking send for FNPSSKDataRequest", or "java.lang.Exception: error at > freenet.node.Peernode.sendSync(PeerNode.java:1517)"... but the rest, > 79.9MB are *only* this one PacketSender ERROR). > > Also, just to further stress the severity of the crashes :) :|, here is > a graph from today, each peak is a restart. (Ignore the rest of the > data.) > > http://dennisn.dyndns.org/guest/pubstuff/freenet-jvm-crashes.png > ___ > Support mailing list > Support at freenetproject.org > http://news.gmane.org/gmane.network.freenet.support > Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support > Or mailto:support-request at freenetproject.org?subject=unsubscribe > > -- 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/20081106/4fcfbf4a/attachment.pgp>
[freenet-support] Freenet 0.7 build 1169
Please forward the below to FMS. Also, if update doesn't work, please run update.sh / update.cmd manually. Thanks. On Thursday 06 November 2008 14:02, Matthew Toseland wrote: > Freenet 0.7 build 1169 is now available. Please update. No massive changes, > but some important bugfixes which hopefully will fix some of the connectivity > problems that have been reported recently: > - Fix announcing to the same node more than once, disconnecting, wasting lots > of announcements. (Thanks sdiz) > - Don't do DNS lookups constantly on the packet sender thread, this caused > havoc with a node with a nonexistent DNS address. > - Logging to try to track down problems with updating. > - Make announcements slightly more tolerant of lag. > - Make the salted-hash datastore the default for new nodes (but not for old > nodes). The salted-hash datastore, written by sdiz, is a massive improvement > in most respects on the old bdbje datastore. It features random replacement > (which has shown in simulations to slightly improve network performance), > uses much less CPU, is much more robust, and uses less in-JVM memory. > However, by default it uses a Bloom filter to minimise disk I/O and improve > performance; this uses 1/2048th of the size of the store, memory mapped to a > disk file; it doesn't count towards the JVM memory limit. You can turn this > off if seeking is not a problem (e.g. on a flash disk). Note that the > salted-hash is only default for nodes that don't already have a datastore > type: old nodes will not be automatically converted, since conversion takes a > huge amount of disk space. > > This is another build with minimal changes - trunk has a number of major > changes, but is not yet ready to release, so this build is based on > 1166/1167/1168 rather than on the current trunk code. > > Please report any bugs you find! Thanks. > -- 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/20081106/e087cc8c/attachment.pgp>
[freenet-support] Freenet 0.7 build 1169
> On Thu, 6 Nov 2008 14:02:40 +, > Matthew Toseland wrote: > > Freenet 0.7 build 1169 is now available. Please update. No massive > changes, but some important bugfixes which hopefully will fix some of > the connectivity problems that have been reported recently: I'm having a LOT of trouble with 1169 (r23352)--the "JVM appears hung" way more often. Actually, although I was able to get it started** the first couple of times, I can't now. It hangs a few minutes after the FEC library is loaded--sometimes at the beginning of a PUBKEY or SSK or CHK-store cleaner, constantly using 100% cpu since being launched. I'm running an Athlon 1.2G, with 384MB ram, and a locally compiled build. Once it outputted a thread dump, though it rarely does this. The one time it did, I had the wrapper max logfile size to 2MB, and the dump filled it with lots of messages like: "java.lang.Thread.State: TIMED_WAITING (parking)" The last few lines were: "VM Thread" prio=10 tid=0x0808 nid=0x145a runnable "VM Periodic Task Thread" prio=10 tid=0x08092000 nid=0x1460 waiting on condition JNI global references: 996 Heap def new generation total 14976K, used 12919K eden space 13312K, 84% used from space 1664K, 99% used to space 1664K, 0% used tenured generation total 198996K, used 147211K the space 198996K, 73% used compacting perm gen total 12288K, used 5663K the space 12288K, 46% used ro space 8192K, 74% used rw space 12288K, 58% used I also remember once while trying to start up, it outputted an error about not being able to open a file (I think it was trying to open libfec) because there were too many files already opened :|. But that only happened once. So yeah, I can't get freenet to start now at all. JVM hangs a few minutes after FEC lib is loaded. ** Also, when I first ran 1169, it forced me to go through the wizard again, and afterwards, although it seemed to remember most of my past settings (node name, logger settings, etc), it didn't remember my datastore size (which i converted to salted hash a few days ago).
[freenet-support] Freenet 0.7 build 1169
> On Thu, 6 Nov 2008 14:42:35 -0500, > Dennis Nezic wrote: > > ., constantly using 100% cpu since being launched. I should also add that my load averages (uptime) were at 80. I've never seen them that high before.
[freenet-support] Freenet 0.7 build 1169
> On Thu, 6 Nov 2008 15:13:11 -0500, > Dennis Nezic wrote: > > I should also add that my load averages (uptime) were at 80. I've > never seen them that high before. Update: I tried it again, except this time I accidentally forgot to give it a seednodes.fref. The first attempt, the same thing happened: after node initialization was completed, after it found update edition 1169, and about three minutes after "Loading FEC lib", the JVM hung. BUT, the wrapper's second attempt immediately afterwards went further! It completed all the 6 cleaners (CHK-cache, CHK-store, SSK-cache, etc), but since there were no seednodes, it didn't connect to anything. BUT it was still consuming all the cpu!--supposedly not doing anything!? At this point I turned the logger back on, and to my horror, it immediately filled to over 70MB, 69.9 of which were errors of the form: (freenet.client.async.SingleBlockInserter, Insert thread for freenet.client.async.SingleFileInserter at 1f4e571(1545), ERROR): Caught java.io.IOException: File already freed encoding data freenet.support.io.TempBucketFactory$TempBucket at 16debe8 java.io.IOException: File already freed at freenet.support.io.BaseFileBucket.getInputStream(BaseFileBucket.java:251) at freenet.support.io.PaddedEphemerallyEncryptedBucket.getInputStream(PaddedEphemerallyEncryptedBucket.java:210) at freenet.support.io.TempBucketFactory$TempBucket$TempBucketInputStream.(TempBucketFactory.java:236) at freenet.support.io.TempBucketFactory$TempBucket.getInputStream(TempBucketFactory.java:220) at freenet.support.io.BucketTools.toByteArray(BucketTools.java:197) at freenet.keys.Key.compress(Key.java:233) at freenet.keys.ClientCHKBlock.encode(ClientCHKBlock.java:141) at freenet.client.async.SingleBlockInserter.innerEncode(SingleBlockInserter.java:83) at freenet.client.async.SingleBlockInserter.encode(SingleBlockInserter.java:117) at freenet.client.async.SingleBlockInserter.getBlock(SingleBlockInserter.java:209) at freenet.client.async.SingleBlockInserter.getURI(SingleBlockInserter.java:249) at freenet.client.async.SingleFileInserter.insert(SingleFileInserter.java:247) at freenet.client.async.SingleFileInserter.access$000(SingleFileInserter.java:37) at freenet.client.async.SingleFileInserter$1.run(SingleFileInserter.java:207) at freenet.support.PooledExecutor$MyThread.run(PooledExecutor.java:190) I wasn't even able to shut the node down at this point... the JVM first.
[freenet-support] Freenet 0.7 build 1169
> On Thu, 6 Nov 2008 21:28:38 -0500, > Dennis Nezic wrote: > > .. BUT it was still consuming all the cpu!--supposedly not > doing anything!? At this point I turned the logger back on, and to my > horror, it immediately filled to over 70MB, 69.9 of which were errors > of the form: > > (freenet.client.async.SingleBlockInserter, > Insert thread for freenet.client.async.SingleFileInserter at 1f4e571 > (1545), ERROR): Caught java.io.IOException: > File already freed encoding data > freenet.support.io.TempBucketFactory$TempBucket at 16debe8 > java.io.IOException: File already freed at > freenet.support.io.BaseFileBucket.getInputStream > (BaseFileBucket.java:251) at If I start with a fresh datastore (if I move my old datastore directory somewhere else), I get a slightly different bunch of error messages, though not nearly as many as before, and all mostly of this type: (freenet.client.async.ClientPutter, Insert thread for freenet.client.async.SingleFileInserter at eaf40c(348), ERROR): addBlock() but set finalized! on freenet.client.async.ClientPutter at 1d87b85 (freenet.client.async.ClientPutter, Insert thread for freenet.client.async.SingleFileInserter at eaf40c(348), ERROR): onTransition: cur=freenet.client.async.MultiPutCompletionCallback at 1de45e2, old=freenet.client.async.SingleFileInserter at eaf40c, new=freenet.client.async.MultiPutCompletionCallback at 122c9df They would usually (not always) occur together, or 1ms apart. There would also somtimes occur a row of: (freenet.client.async.ClientPutter, Pooled thread awaiting work @70, ERROR): addBlock() but set finalized! on freenet.client.async.ClientPutter at 1d87b85 And I was able to easily shut down the node in this case :|. (No JVM hangs :\.)
[freenet-support] Freenet 0.7 build 1169
Hrm.. sorry.. it seems to be working now that I'm using r23350. I was having all my problems with r23352--which I could have sworn was referenced in freenet-stable-latest.jar.url this morning :|. Anywho, sorry for the noise.
[freenet-support] Freenet 0.7 build 1169
Freenet 0.7 build 1169 is now available. Please update. No massive changes, but some important bugfixes which hopefully will fix some of the connectivity problems that have been reported recently: - Fix announcing to the same node more than once, disconnecting, wasting lots of announcements. (Thanks sdiz) - Don't do DNS lookups constantly on the packet sender thread, this caused havoc with a node with a nonexistent DNS address. - Logging to try to track down problems with updating. - Make announcements slightly more tolerant of lag. - Make the salted-hash datastore the default for new nodes (but not for old nodes). The salted-hash datastore, written by sdiz, is a massive improvement in most respects on the old bdbje datastore. It features random replacement (which has shown in simulations to slightly improve network performance), uses much less CPU, is much more robust, and uses less in-JVM memory. However, by default it uses a Bloom filter to minimise disk I/O and improve performance; this uses 1/2048th of the size of the store, memory mapped to a disk file; it doesn't count towards the JVM memory limit. You can turn this off if seeking is not a problem (e.g. on a flash disk). Note that the salted-hash is only default for nodes that don't already have a datastore type: old nodes will not be automatically converted, since conversion takes a huge amount of disk space. This is another build with minimal changes - trunk has a number of major changes, but is not yet ready to release, so this build is based on 1166/1167/1168 rather than on the current trunk code. Please report any bugs you find! Thanks. pgpEm6tzaitXc.pgp Description: PGP signature ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Node won't connect to peers 1167/68
Okay, the PacketSender resent-and-not-acked problem is probably caused by a bug fixed in 1169. Unfortunately we will have to wait for most of the network to update before this will make a big impact. On Thursday 06 November 2008 02:43, Dennis Nezic wrote: > > On Wed, 5 Nov 2008 22:00:52 +, > > Matthew Toseland <[EMAIL PROTECTED]> wrote: > > > > It could very well be. Lack of memory resulting in constant garbage > > collection could easily cause a timeout. > > I really don't think it's low memory. I increased > wrapper.java.maxmemory to 300, and even though only about 180MB is > being used (only 250MB is allocated.. ie. 50MB hasn't been touched > yet), it still crashes. (Also, varying maxmemory between 180 and 300 > doesn't really affect the interval between crashes.) > > Just to further stress the severity of the PacketSender "packet > not acked" errors that I'm getting, my freenet-previous.log is 80MB! > (0.1MB of which is another RequestSender "Waited too long for a > blocking send for FNPSSKDataRequest", or "java.lang.Exception: error at > freenet.node.Peernode.sendSync(PeerNode.java:1517)"... but the rest, > 79.9MB are *only* this one PacketSender ERROR). > > Also, just to further stress the severity of the crashes :) :|, here is > a graph from today, each peak is a restart. (Ignore the rest of the > data.) > > http://dennisn.dyndns.org/guest/pubstuff/freenet-jvm-crashes.png > ___ > Support mailing list > Support@freenetproject.org > http://news.gmane.org/gmane.network.freenet.support > Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support > Or mailto:[EMAIL PROTECTED] > > pgp3y3bvY6TWt.pgp Description: PGP signature ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Freenet 0.7 build 1169
Please forward the below to FMS. Also, if update doesn't work, please run update.sh / update.cmd manually. Thanks. On Thursday 06 November 2008 14:02, Matthew Toseland wrote: > Freenet 0.7 build 1169 is now available. Please update. No massive changes, > but some important bugfixes which hopefully will fix some of the connectivity > problems that have been reported recently: > - Fix announcing to the same node more than once, disconnecting, wasting lots > of announcements. (Thanks sdiz) > - Don't do DNS lookups constantly on the packet sender thread, this caused > havoc with a node with a nonexistent DNS address. > - Logging to try to track down problems with updating. > - Make announcements slightly more tolerant of lag. > - Make the salted-hash datastore the default for new nodes (but not for old > nodes). The salted-hash datastore, written by sdiz, is a massive improvement > in most respects on the old bdbje datastore. It features random replacement > (which has shown in simulations to slightly improve network performance), > uses much less CPU, is much more robust, and uses less in-JVM memory. > However, by default it uses a Bloom filter to minimise disk I/O and improve > performance; this uses 1/2048th of the size of the store, memory mapped to a > disk file; it doesn't count towards the JVM memory limit. You can turn this > off if seeking is not a problem (e.g. on a flash disk). Note that the > salted-hash is only default for nodes that don't already have a datastore > type: old nodes will not be automatically converted, since conversion takes a > huge amount of disk space. > > This is another build with minimal changes - trunk has a number of major > changes, but is not yet ready to release, so this build is based on > 1166/1167/1168 rather than on the current trunk code. > > Please report any bugs you find! Thanks. > pgpBgivxbKiXP.pgp Description: PGP signature ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Freenet 0.7 build 1169
> On Thu, 6 Nov 2008 14:02:40 +, > Matthew Toseland <[EMAIL PROTECTED]> wrote: > > Freenet 0.7 build 1169 is now available. Please update. No massive > changes, but some important bugfixes which hopefully will fix some of > the connectivity problems that have been reported recently: I'm having a LOT of trouble with 1169 (r23352)--the "JVM appears hung" way more often. Actually, although I was able to get it started** the first couple of times, I can't now. It hangs a few minutes after the FEC library is loaded--sometimes at the beginning of a PUBKEY or SSK or CHK-store cleaner, constantly using 100% cpu since being launched. I'm running an Athlon 1.2G, with 384MB ram, and a locally compiled build. Once it outputted a thread dump, though it rarely does this. The one time it did, I had the wrapper max logfile size to 2MB, and the dump filled it with lots of messages like: "java.lang.Thread.State: TIMED_WAITING (parking)" The last few lines were: "VM Thread" prio=10 tid=0x0808 nid=0x145a runnable "VM Periodic Task Thread" prio=10 tid=0x08092000 nid=0x1460 waiting on condition JNI global references: 996 Heap def new generation total 14976K, used 12919K eden space 13312K, 84% used from space 1664K, 99% used to space 1664K, 0% used tenured generation total 198996K, used 147211K the space 198996K, 73% used compacting perm gen total 12288K, used 5663K the space 12288K, 46% used ro space 8192K, 74% used rw space 12288K, 58% used I also remember once while trying to start up, it outputted an error about not being able to open a file (I think it was trying to open libfec) because there were too many files already opened :|. But that only happened once. So yeah, I can't get freenet to start now at all. JVM hangs a few minutes after FEC lib is loaded. ** Also, when I first ran 1169, it forced me to go through the wizard again, and afterwards, although it seemed to remember most of my past settings (node name, logger settings, etc), it didn't remember my datastore size (which i converted to salted hash a few days ago). ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Freenet 0.7 build 1169
> On Thu, 6 Nov 2008 14:42:35 -0500, > Dennis Nezic <[EMAIL PROTECTED]> wrote: > > ., constantly using 100% cpu since being launched. I should also add that my load averages (uptime) were at 80. I've never seen them that high before. ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
[freenet-support] Node can't find native BigInteger library
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Just noticed that, for some reason, my node doesn't pick up native BigInteger library: ... Initializing Node using Freenet Build #1169 r23350M and freenet-ext Build #24 r23199 with Sun Microsystems Inc. JVM version 1.6.0_06 running on amd64 Windows 2003 5.2 ... NOTICE: Resource name [net/i2p/util/jbigi-windows-x86_64.dll] was not found Library net/i2p/util/jbigi-windows-none.dll is not appropriate for this system. INFO: Native BigInteger library jbigi not loaded - using pure java ... Anything I can do to get the node to load a native library? On a related topic, Windows x86-64 native binary for Onion FEC isn't present as well. Are there plans to include it in the foreseeable future? Regards, Victor Denisov. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFJE4XM1O5++4rTuI0RArMFAKCCUZ6LotMJ2Akz3Gpg3jZc1qhUDQCgnMlf ZJWBolsNbsl9afK5ICcD4yE= =MYhB -END PGP SIGNATURE- ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Node can't find native BigInteger library
On Friday 07 November 2008 00:03, Victor Denisov wrote: > Just noticed that, for some reason, my node doesn't pick up native > BigInteger library: > > ... > > Initializing Node using Freenet Build #1169 r23350M and freenet-ext > Build #24 r23199 with Sun Microsystems Inc. JVM version 1.6.0_06 running > on amd64 Windows 2003 5.2 > > ... > > NOTICE: Resource name [net/i2p/util/jbigi-windows-x86_64.dll] was not found > Library net/i2p/util/jbigi-windows-none.dll is not appropriate for this > system. > INFO: Native BigInteger library jbigi not loaded - using pure java > > ... > > Anything I can do to get the node to load a native library? Dunno. We do have them for linux. > > On a related topic, Windows x86-64 native binary for Onion FEC isn't > present as well. Are there plans to include it in the foreseeable future? That's a bigger problem, the code isn't 64-bit clean. Nextgens tried to port it but didn't succeed. pgpqt5iS0tNcE.pgp Description: PGP signature ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Freenet 0.7 build 1169
> On Thu, 6 Nov 2008 15:13:11 -0500, > Dennis Nezic <[EMAIL PROTECTED]> wrote: > > I should also add that my load averages (uptime) were at 80. I've > never seen them that high before. Update: I tried it again, except this time I accidentally forgot to give it a seednodes.fref. The first attempt, the same thing happened: after node initialization was completed, after it found update edition 1169, and about three minutes after "Loading FEC lib", the JVM hung. BUT, the wrapper's second attempt immediately afterwards went further! It completed all the 6 cleaners (CHK-cache, CHK-store, SSK-cache, etc), but since there were no seednodes, it didn't connect to anything. BUT it was still consuming all the cpu!--supposedly not doing anything!? At this point I turned the logger back on, and to my horror, it immediately filled to over 70MB, 69.9 of which were errors of the form: (freenet.client.async.SingleBlockInserter, Insert thread for [EMAIL PROTECTED](1545), ERROR): Caught java.io.IOException: File already freed encoding data [EMAIL PROTECTED] java.io.IOException: File already freed at freenet.support.io.BaseFileBucket.getInputStream(BaseFileBucket.java:251) at freenet.support.io.PaddedEphemerallyEncryptedBucket.getInputStream(PaddedEphemerallyEncryptedBucket.java:210) at freenet.support.io.TempBucketFactory$TempBucket$TempBucketInputStream.(TempBucketFactory.java:236) at freenet.support.io.TempBucketFactory$TempBucket.getInputStream(TempBucketFactory.java:220) at freenet.support.io.BucketTools.toByteArray(BucketTools.java:197) at freenet.keys.Key.compress(Key.java:233) at freenet.keys.ClientCHKBlock.encode(ClientCHKBlock.java:141) at freenet.client.async.SingleBlockInserter.innerEncode(SingleBlockInserter.java:83) at freenet.client.async.SingleBlockInserter.encode(SingleBlockInserter.java:117) at freenet.client.async.SingleBlockInserter.getBlock(SingleBlockInserter.java:209) at freenet.client.async.SingleBlockInserter.getURI(SingleBlockInserter.java:249) at freenet.client.async.SingleFileInserter.insert(SingleFileInserter.java:247) at freenet.client.async.SingleFileInserter.access$000(SingleFileInserter.java:37) at freenet.client.async.SingleFileInserter$1.run(SingleFileInserter.java:207) at freenet.support.PooledExecutor$MyThread.run(PooledExecutor.java:190) I wasn't even able to shut the node down at this point... the JVM first. ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Freenet 0.7 build 1169
> On Thu, 6 Nov 2008 21:28:38 -0500, > Dennis Nezic <[EMAIL PROTECTED]> wrote: > > .. BUT it was still consuming all the cpu!--supposedly not > doing anything!? At this point I turned the logger back on, and to my > horror, it immediately filled to over 70MB, 69.9 of which were errors > of the form: > > (freenet.client.async.SingleBlockInserter, > Insert thread for [EMAIL PROTECTED] > (1545), ERROR): Caught java.io.IOException: > File already freed encoding data > [EMAIL PROTECTED] > java.io.IOException: File already freed at > freenet.support.io.BaseFileBucket.getInputStream > (BaseFileBucket.java:251) at If I start with a fresh datastore (if I move my old datastore directory somewhere else), I get a slightly different bunch of error messages, though not nearly as many as before, and all mostly of this type: (freenet.client.async.ClientPutter, Insert thread for [EMAIL PROTECTED](348), ERROR): addBlock() but set finalized! on [EMAIL PROTECTED] (freenet.client.async.ClientPutter, Insert thread for [EMAIL PROTECTED](348), ERROR): onTransition: [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] They would usually (not always) occur together, or 1ms apart. There would also somtimes occur a row of: (freenet.client.async.ClientPutter, Pooled thread awaiting work @70, ERROR): addBlock() but set finalized! on [EMAIL PROTECTED] And I was able to easily shut down the node in this case :|. (No JVM hangs :\.) ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
Re: [freenet-support] Freenet 0.7 build 1169
Hrm.. sorry.. it seems to be working now that I'm using r23350. I was having all my problems with r23352--which I could have sworn was referenced in freenet-stable-latest.jar.url this morning :|. Anywho, sorry for the noise. ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
