Scheduler and DB versions are now both called getNextRetrieveJobsFailedBatch()

* archive_and_retrieve_failure - fails archive, handles the failure report
* archive_and_retrieve_report_failure - fails archive and also fails the report

In both cases, the job ends up on the failed queue.

Default CTA Frontend log level is changed to INFO. To set it back to
debug level, add this line to /etc/cta/cta-frontend-xrootd.conf:

cta.log.level DEBUG

Valid log levels are EMERG, ALERT, CRIT, ERR, WARNING,
NOTICE (==USERERR), INFO, DEBUG

Upgraded to EOS 4.4.10-20181106193022gitf4dd861 in order to get the propagation of WFE errors and the correct use of locking around fmd for mTapeAwareGc.fileReplicaCommitted()

Upgraded to EOS 4.4.10-20181106193022gitf4dd861 in order to get the new tape aware GC configuration system

Renamed rdbms::AutocommitMode::ON and rdbms::AutocommitMode::OFF to rdbms::AutocommitMode::AUTOCOMMIT_ON and rdbms::AutocommitMode::OFF respectively

Moved OracleCatalogue::deleteArchiveFile to RdbmsCatalogue because Oracle, MySQL and PostgreSQL support 'SELECT FOR UPDATE'.  SQLite is the odd one out in that it does not provide support.