Why does the refresh or compare job file get deleted from the folder $HVR_CONFIG/job<hubname><channelname>?
This can happen if the user creates multiple refresh or compare jobs for the same channel with the default task name. The default task name identifier for refresh is 'refr' and for compare it is 'cmp'. So by default HVR will create the job name as <channelname>-refr-<source_location>-<target_location> for a refresh job and <channelname>-cmp-<source_location>-<target_location> for compare job.
When a job is scheduled in HVR, a file is created in folder $HVR_CONFIG job <hubname> <channelname>. For example, let's say you have scheduled a refresh job for table 'orders' in channel ch1 using default task naming, replicating from Oracle source on windows (ora) to Oracle on linux vm (vmtgt). The job will be created as 'chn1-refr-ora-vmtgt'. If subsequent jobs are then scheduled for the same channel using default task naming (refror cmp), the already-existing job will be replaced by the newly scheduled job of the same name. This makes it appear as if the old job is getting deleted.
Create refresh or compare jobs with user-defined task names. For example, when creating a job to refresh the 'orders' table, supply a unique task name such as refrords. This will create the job as ch1-refrords-ora-vmtgt. Similarly when creating a job to refresh the 'history' table, supply a task name such as refrhist. This will create the job as chn1-refrhist-ora-vmtgt.