saces / fred Goto Github PK

seems to be a rare race condition

INFO | jvm 1 | 2011/09/02 16:08:55 | [Sep 02, 2011 14:08:55:589] [ERROR](freenet.node.fcp.ClientPutDir, Client database access thread%2813%29, ERROR): No generated URI in onSuccess() for freenet.node.fcp.ClientPutDir@284a6c0 from freenet.client.async.SimpleManifestPutter@336f356c
INFO | jvm 1 | 2011/09/02 16:08:55 | [Sep 02, 2011 14:08:55:599] [ERROR](freenet.node.NodeClientCore$DBJobWrapper, Client database access thread%2813%29, ERROR): Failed to run database job freenet.client.async.PersistentChosenRequest$1@786d2928 : caught java.lang.NullPointerException
INFO | jvm 1 | 2011/09/02 16:08:55 | java.lang.NullPointerException
INFO | jvm 1 | 2011/09/02 16:08:55 | at java.util.Hashtable.get(Hashtable.java:334)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.support.MultiValueTable.put(MultiValueTable.java:33)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.node.fcp.RequestStatusCache.finishedUpload(RequestStatusCache.java:86)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.node.fcp.FCPClient.finishedClientRequest(FCPClient.java:171)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.node.fcp.ClientRequest.finish(ClientRequest.java:276)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.node.fcp.ClientPutBase.onSuccess(ClientPutBase.java:188)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.node.fcp.ClientPutDir.onSuccess(ClientPutDir.java:368)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.client.async.SimpleManifestPutter.complete(SimpleManifestPutter.java:1277)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.client.async.SimpleManifestPutter.onSuccess(SimpleManifestPutter.java:1553)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.client.async.MultiPutCompletionCallback.complete(MultiPutCompletionCallback.java:185)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.client.async.MultiPutCompletionCallback.onSuccess(MultiPutCompletionCallback.java:102)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.client.async.SingleBlockInserter.onSuccess(SingleBlockInserter.java:475)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.client.async.PersistentChosenRequest.finish(PersistentChosenRequest.java:224)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.client.async.PersistentChosenRequest.access$0(PersistentChosenRequest.java:148)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.client.async.PersistentChosenRequest$1.run(PersistentChosenRequest.java:137)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.node.NodeClientCore$DBJobWrapper.run(NodeClientCore.java:1822)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.support.PrioritizedSerialExecutor$Runner.run(PrioritizedSerialExecutor.java:99)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.support.PooledExecutor$MyThread.realRun(PooledExecutor.java:233)
INFO | jvm 1 | 2011/09/02 16:08:55 | at freenet.support.io.NativeThread.run(NativeThread.java:130)

The BaseManifestPutter is still leaking persistent objects.
freenet.client.async.ContainerInserter should be reviewed/fixed first, the whole thing is left in db.

The current 'export' fetches the block, decodes it (good for verification), then encode
it again and add the new encoded to the output blob.
The compression codec it was compressed with is not passed out, recompressing with best may give an other blob then the original blob was, especially if a weak compression codec was explicit given at insert time..