Mantis - Resin
Viewing Issue Advanced Details
1911 minor always 07-30-07 09:07 08-28-07 19:28
ferg  
ferg  
normal  
closed  
fixed  
none    
none 3.1.3  
0001911: cluster sessions: block 0 is reserved
(rep by Ethan Larson)

We just completed the upgrade to 3.1.2 from 3.1.1, which included a
switch to clusterstore from jdbcstore as well. Things look good, except
when we shut one of the nodes down, some of the other nodes in the
cluster will sometimes throw this exception:

[11:57:01.935] java.lang.IllegalStateException: Block 0 is reserved.
[11:57:01.935] at
com.caucho.db.store.BlockManager.stateError(BlockManager.java:329)
[11:57:01.935] at
com.caucho.db.store.BlockManager.getBlock(BlockManager.java:240)
[11:57:01.935] at com.caucho.db.store.Store.readBlock(Store.java:466)
[11:57:01.935] at
com.caucho.db.store.Store.readMiniFragment(Store.java:1319)
[11:57:01.935] at com.caucho.db.store.Inode.read(Inode.java:302)
[11:57:01.935] at
com.caucho.db.store.BlobInputStream.read(BlobInputStream.java:114)
[11:57:01.935] at
com.caucho.server.cluster.ClusterStore.dumpObjects(ClusterStore.java:1231)
[11:57:01.935] at
com.caucho.server.hmux.HmuxClusterRequest.dumpObjects(HmuxClusterRequest.java:391)
[11:57:01.935] at
com.caucho.server.hmux.HmuxClusterRequest.handleRequest(HmuxClusterRequest.java:178)
[11:57:01.935] at
com.caucho.server.hmux.HmuxRequest.scanHeaders(HmuxRequest.java:641)
[11:57:01.935] at
com.caucho.server.hmux.HmuxRequest.handleRequest(HmuxRequest.java:354)
[11:57:01.935] at
com.caucho.server.port.TcpConnection.run(TcpConnection.java:481)
[11:57:01.935] at
com.caucho.util.ThreadPool$Item.runTasks(ThreadPool.java:685)
[11:57:01.935] at com.caucho.util.ThreadPool$Item.run(ThreadPool.java:607)
[11:57:01.935] at java.lang.Thread.run(Thread.java:619)

Is this something to worry about? Should I file a bug? I found a
similar bug in mantis here: http://bugs.caucho.com/view.php?id=4 [^]

Notes
(0002246)
ferg   
08-28-07 19:28   
SelectResult needed to be updated to correct size of TempBuffer.SIZE