Correct the handling of an EQQJTARC file full (SB37 abend) to ensure that job tracking records are not lost, and jobs are not submitted twice

Currently, if the EQQJTARC file becomes full, immediately the zWS  JT ARCHIVER subtask fails (message EQQZ045E SUBTASK JT LOG ARCHIVER  ENDED UNEXPECTEDLY) however it is still possible to write job tracking (JT) records to the JT files which are not already full and in need of being archived (i.e. EQQJT01 is copied to EQQJTARC).  This continues until ALL the JT files are full and need to be archived.  At this point the normal mode manager (NMM) subtask files and zWS can no longer track jobs.  However, if after the JTARC file becomes full and before the NMM subtask fails (which could be hours or even days, depending on the activity within the zWS controller, and how larger the EQQJT0x  files are) a CP batch job like CP EXTEND is started, the job will hang indefinitely as it needs to copy the EQQJTARC file to the Tracklog (EQQTROUT) and then clear the EQQJTARC file.  In the customer  case recorded here, this actually happened- the CP EXTEND task was hanging for hours and eventually the NMM subtask failed. A recovery of the JTARC file was done, but it was too late- job tracking records had been lost. As a result of this, when the controller was restarted after fixing the EQQJTARC file, jobs which had previously been submitted and completed, were submitted again.  Depending on the type of job this could have very serious consequences.