Problem | An empty logfile was given to the ObjectManager so it was cold started. |
User response | None, unless you expected the ObjectManager to warm start. In which case, investigate why the log file was empty. |
Problem | A valid logfile was given to the ObjectManager which it was able to restart from. |
User response | None, unless you expected the ObjectManager to cold start. In which case, investigate why the log file was not empty. |
Problem | The ObjectManager was requested to shutdown. |
User response | None, unless you did not expect the ObjectManager to shut down. |
Problem | The ObjectManager was requested to shutdown without performing a final checkpoint. |
User response | None, unless you did not expect the ObjectManager to shut down without performing a final checkpoint. |
Problem | The ObjectManager was short of log file space after it had completed a checkpoint, it backed out the largest transaction in the log in order to release some space. |
User response | You are trying to do too much transactional work for the size of log file you have configured. Consider whether your transactions are attempting to do too many updates, or whether you are trying to update too many large objects, or whether you are trying to run too many transactions at the same time, or whether your log file is too small and should be made larger. |
Problem | The application lost its reference to a transaction and would not be able to complete it so the ObjectManager will back out the transaction in order to free up its resources. |
User response | Your application lost its reference to a transaction. If the application terminated abnormally there us nothing further to do. If however, the application lost its reference without terminating abnormally this could be considered to be a logic error in the application. |
Problem | If there are active transactions at shutdown. ObjectStores with strategy STRATEGY_SAVE_ONLY_ON_SHUTDOWN cannot safely be closed as this might lead to their saving inconsistent data. The ObjectStore is left unchanged. |
User response | Examine why there are active transactions, this might be because you are using an external transaction manager. If you cannot avoid having active transactions at shutdown then use an ObjectStore with a different storage strategy. |
Problem | The file system could not allocate the required space. Attempts to allocate objects in the ObjectStore will fail. |
User response | Examine the exception and determine why the space could not be allocated in thefile system. |
Problem | A java.lang.ClassNotFoundException was caught while trying to deserialize a ManagedObject. |
User response | Look at the class that could not be found and make it available to the ObjectManagers class loader. |
Problem | ObjectStore names must be unique within an ObjectManager. |
User response | Use a different name for the ObjectStore. |
Problem | An operation was requested on an object in an invalid state, the operation fails the object is transitioned to an error state. |
User response | This is an internal logic error in the ObjectManager. No further operations on the Object can succeed, contact IBM support. |
Problem | An IOException was caught by the ObjectManager, the IO operation could not be safely retried and was abandoned. |
User response | Investigate and fix the cause of the underlying IOException. |
Problem | An IOException was caught by the ObjectManager, the IO operation was abandoned, but could be safely retried by retrying the ObjectManager API, once the underlying cause is fixed. |
User response | Investigate and fix the cause of the underlying IOException, retry the operation. |
Problem | An ObjectStore was constructed using an invalid name. The ObjectStore is not created. |
User response | Examine the name to determine why it is not valid. |
Problem | A set of statistics which is not recognized was requested. |
User response | Correct the name of the statistics set requested. |
Problem | An ObjectStore was asked to store an invalid ManagedObject. The ManagedObject is not stored. |
User response | Examine the ManagedObject to be stored and determine why it is invalid. For example, it might be null. |
Problem | An internal error has occurred in the ObjectStore. |
User response | Report the problem to IBM service. |
Problem | Only one token of each identity (objectStoreIdentifier and storedObjectIdentifier; store may exist, but the new Token and an existing one share the same identity. |
User response | The ObjectStore may be corrupt, contact IBM support. |
Problem | An invalid operation was attempted on the object for the state it is in. The operation is rejected and the Object is unchanged. |
User response | Investigate the state of the Object and why the invalid operation is being attempted. |
Problem | The ObjectManager has discovered a new and existing transaction that have the same LUWID. This is an internal error and should not occur. |
User response | Collect the available information and contact IBM support. |
Problem | This is an error in the ObjectManager and should not occur. No further actions are allowed on the Transaction. |
User response | Gather any FFDC and other related information and contact IBM support. |
Problem | An exception was thrown and caught by the ObjectManager, the ObjectManager was not able to take any reasonable action in response to the exception. |
User response | Investigate the cause of the underlying exception. |
Problem | An attempt was made to unlock or replace a ManagedObject under a different transaction to the one that locked it. The attempt is rejected. |
User response | Examine the logic of the application using the ObjectManager Transaction.unlock(ManagedObject) interface to see why the two transactions are different. |
Problem | An exception was caught trying to open the named log file. The ObjectManager will not start. |
User response | Check that the log file name is valid, and that the ObjectManager has the appropriate permissions to write it. |
Problem | The ObjectManager has detected the end of the input log file. |
User response | None if this is raised as a result of reading the log to the end at restart of the ObjectManager. |
Problem | The ObjectManager was performing its recovery and found a log record with a type that it did not recognize, the ObjectManager is not instantiated. |
User response | Examine the log file that the ObjectManager is using and determine of it is corrupt. If it has not been corrupted contact IBM support. |
Problem | The ObjectManager was asked to locate an ObjectStore that was not registered with it. |
User response | The ObjectStore may have been removed from the ObjectManager. Recreate the ObjectStore. |
Problem | In internal error has occurred and the invalid condition found. |
User response | The action is terminated. An internal error has occurred, gather the available data and report the error to IBM service. |
Problem | The ObjectManager tried to obtain an exclusive lock on the log file but it was already being used. |
User response | See if another ObjectManager or another program is trying to use the same log file. |
Problem | The transaction log file contained data that the ObjectManager could not interpret. |
User response | Check that the log file is the correct one. If it is, contact IBM support. |
Problem | A transaction may only have one XID and it cannot be modified once set. |
User response | Examine the logic of the application and determine why it is trying to change the XID of a transaction that already has one. |
Problem | An XID can have a maximum length of java.lang.Short.MAX_VALUE. |
User response | Use a shorter XID length. |
Problem | An attempt was made to use an interface that is disabled. |
User response | Stop using the interface, or switch to a version of the ObjectManager that supports it. |
Problem | An attempt was made to write more data to the log than the available file space allowed. The attempt is rejected, the transaction state is unchanged. |
User response | Either, reduce the number of objects you update within your transactions, reduce the number of transactions that are simultaneously active, reduce the size of the ManagedObjects you update or increase the size of your log file. |
Problem | A type flag on log record part was not recognized. |
User response | If the log file has not been corrupted contact IBM support. |
Problem | The log file ended before a checkpoint end record was found. Log files should contain at least one checkpoint start and at least one checkpoint end logrecord. |
User response | If the log file has not been corrupted contact IBM support. |
Problem | The ObjectManager was asked to retrieve a ManagedObject which should have been in memory but was not. |
User response | Contact IBM support. |
Problem | The size of a serialized ManagedObject exceeded the maximum expected. |
User response | This should not occur, gather all available information and contact IBM support. |
Problem | The ObjectManager was asked to reduce the size of it's transaction log. The requested size was too small to contain the existing log data. The new size is below the limit which will currently allow the ObjectManager to function without continually checkpointing. |
User response | Either make a larger request or reduce the load on the log by writing fewer or smaller ManagedObjects. |
Problem | The ObjectManager was asked to delete a collection ( Map or List ) that was not empty,it contained some entries that were not deleted or were deleted by a transaction other than the one trying to delete the collection. |
User response | Examine the logic of the application and discover why it is trying to delete the collection while it is not empty. |
Problem | The ObjectManager was asked to start more transactions than it has been configured for. The new transaction is not started. |
User response | Examine the logic of the application and discover why it tried to start too many transactions. Possibly it is starting them but not committing them or backing them out. |
Problem | The ObjectManager was asked to start more transactions than it can support. Usually this has been reduced so that checkpoints can complete before the log file fills. The new transaction is not started. |
User response | Either increase the size of the log file or start fewer transactions. |
Problem | The ObjectStore found that it was already locked when it tried to take an exclusive lock. |
User response | See if the ObjectStore is being used by another ObjectManager of if its file is being used by another program. |
Problem | The log file does not end normally, the physical file appears to have been truncated. |
User response | Determine of the log file has been damaged, restore the original undamaged file. |
Problem | The ObjectManager was asked to store a reference to a named ManagedObject but there were no restartable ObjectStores in which to store it. |
User response | Consider whether you need to use a named ManagedObject and if so, create a restartable ObjectStore. |
Problem | An attempt was made to add a unique key which already existed in the map and was not in a suitable state for replacement. |
User response | Either request addition of a duplicate key or find out why the application is trying to replace an existing key. |
Problem | The ObjectManager was asked to reduce the size of a store file to a length that could not contain the existing ManagedObjects in the file. |
User response | Make a request that is for more space than the currently used size. |
Problem | An nio Exception was caught by the ObjectManager, the IO operation could not be safely retried and was abandoned. |
User response | Investigate and fix the cause of the underlying Exception. |
Problem | An ObjectStore was already full when a new allocate request was made. The allocation was not made. |
User response | Decide if you should make the ObjectStore larger or if you should remove some existing objects, or if you are trying to store more or larger objects than you had planned. |
Problem | The ObjectManager constructor was passed a log file type that was not recognized as one of the LOG_FILE_TYPE_XXXX types. The ObjectManager is not instantiated. |
User response | Change the constructor invocation to use a recognized log file type. |
Problem | A list was asked to create a sublist delimited by entries it did not contain. The sublist was not created. |
User response | Look at the application code to determine why the entry was not in the list. |
Problem | An unrecognize signature was found in a ManagedObject, the deserialization was abandoned. |
User response | Contact IBM support. |
Problem | ObjectStores can only be constructed with valid storage strategies. |
User response | Use a valid storage strategy to construct the ObjectStore. |
Problem | ObjectStore identifiers must be unique within an ObjectManager. |
User response | Add existing ObjectStores to the ObjectManagerbefore creating new ones, this allows the ObjectManager to avoid reusing the identifier. |
Problem | For data protected by Guard bytes, a mismatch in the value of the Guard bytes has been found. |
User response | Contact IBM support. |
Problem | The ObjectStore has a corrupt header and cannot be processed. |
User response | Determine how the file was corrupted and restore a backup. |
Problem | The log file has a corrupt header and cannot be processed. |
User response | Determine how the file was corrupted and restore a backup. |
Problem | If the message does not give sufficient information, check previous messages for further help. |
User response | See previous messages for further information. |