- August 29, 2017 at 4:57 pm #10989ggoodrichKeymaster
Background: All replication jobs (capture and integration) are controlled by the HVR scheduler process which runs on the hub machine. The HVR Scheduler gives single-point-of-control; all the log files are in one place and we can see the exact status of everything in the GUI. Each time the HVR Scheduler ‘runs’ a job, that job connects to the remote machine; it starts a HVR agent executable on that machine, and then does all the processing for a “replication cycle”.
Normally these replication jobs run continuously; this means that when one replication cycle is done it goes right on to do the next cycle. But in some cases the replication jobs should not run continuously; only a specific times.
One approach is to use an action such as Scheduling /CaptureStartTimes. This can instruct to run a job at specific moments. There are lots of possibilities, e.g. once every day at 4pm.
Another approach is to define the action Scheduling /CaptureOnceOnTrigger. This means that the job will only run when command hvrtrigger is run on the remote machine. This can be useful if a remote location is often offline, and connections should only be done sometimes. This hvrtrigger command connects to the scheduler process and forces the scheduler so startup the replication job. The job then makes a new connection back from the hub to HVR’s port number on the remote machine and does the necessary replication. While the replication job is doing its cycle, the hvrscheduler keeps the hvrtrigger command (on the remote machine) ‘informed’ about progress; it sends it a copy of any error messages (but the scheduler still writes the message to the central log file), and it terminates the hvrtrigger session once the replication job is complete.
Note that different scheduling can be combined. For example, a channel could do continuous replication to some locations, whereas other locations in the same channel could be both triggered ‘manually’ (via hvrtrigger ) and once each midnight (via Scheduling /CaptureStartTimes).
To run a job outside of the hvrscheduler (i.e. interactively on the command line) one can use hvrtrigger -i <job-name>. This will run 1 cycle for this job. Please note that the output for this run will not be re-directed to the HVR log files (because that is done by the hvrscheduler). It will appear in the terminal window instead.
See hvrtrigger for more information.
- The forum ‘Expert Notes’ is closed to new topics and replies.