Problem:
JCL jobs end with condition code 0000, and their job logs indicate completion, but on the ESMAC screen, the jobs are in a "waiting" state, the JCL initiator SEPs are still running the jobs, and the jobs still hold resource locks on the datasets that were involved. The locks prevent subsequent related jobs from running. A production region is halted because of these "waiting" jobs.
Resolution:
It turned out that a JCL User Exit had recently been modified, and newly-introduced faulty logic within the User Exit prevented the Exit from concluding, and this was the overall reason jobs were stuck in "waiting" status.
Where the documentation describes JCL User Exits, it explains that such exits are capable of logging events to the MVS data set MFMVS.SYSLOG if the User Exit sets appropriate logging flags. To prevent problems of this type, users should include code in User Exits to log events to aid in debugging, and to react to all situations without potentially causing a hang.
#EnterpriseDeveloper
#MFDS
