View on GitHub

# Quelea

## Troubleshooting

### Index

Here is how the error looks:

This error occurs when a previous run of the experiment did not perform proper cleanup before terminating, resulting on zombie processes:

The fix is to kill such processes.

Here is how the error looks like:

This error occurs when a previous run of the experiment failed to delete all Cassandra column families that were created during the run. Straightforward fix is to run the same binary with --kind Drop option:

Another way is to use cqlsh to manually drop tables:

\$ cqlsh
cqlsh> USE Codeec;
cqlsh> DROP TABLE bankaccount;
cqlsh> DROP TABLE bankaccount_lock;
cqlsh> DROP TABLE bankaccount_gc;


### Unconfigured column family error

Here is how the error looks like:

This error occurs when a table being used by the current run of the experiment was dropped in a concurrent run. The fix is to simply rerun the experiment.

### One of the client threads hangs

For example, consider the following sample run:

Even though numThreads option is set to 4, only three threads ever make progress. The other thread simply hangs. This error occurs frequently on a VM that is severely resource constrained. The exact cause of the bug is not yet known. Increasing the resources allocated to the VM will prevent this error from occuring. If this is not possible, we suggest that you rerun the experiment until you generate an execution where all threads make progress.

### Program becomes unresponsive

This is a more severe manifestation of the above bug, where none of the threads make progress. As with the previous case, we suggest that you rerun the experiment until you generate an execution where all threads make progress.