FIX BUG: When requesting a specific job, the job was re-run until until the workers had run jobs for at least MIN_BATCH_TIME

as change it to use ensembl-analysis breaks other dependencies.

ensembl-pipeline as the later has been declared obsolete and modules have
been removed from the repository.

Change in the monitor table: analyses -> analysis to improve auto-completion in mysql client for all the analysis_* tables.

replaced the remove/delete analysis -- shouldnt be using that unless you know what you are doing -- not advertised in the help anymore -- Albert Vilella and Michael Schuster

the deletion of the method is done in the right place now - still, be careful about using this method - Albert Vilella and Michael Schuster

adding a remove_analysis_id method that will DELETE FROM analysis, analysis_stats, and analysis_job table WHERE analysis_id equals the number given

Change: performs global cleanup after a worker dies unless no global cleanup has been explicitly requested

Now takes into account lsf_pending jobs if any, before deciding to do a hard sync. Added specific exit message when beekeeper is running a specific analysis that is completed

very rare circumstances, caused error messages to be sent to space