When is "moving_data.total_written_bytes" reset?

The “moving_data.total_written_bytes” parameter is reset to 0 from time to time. At what time or under what conditions is it reset?

Below is a screenshot of the parameter over the last 4 days in our cluster. The reset was not fixed at specific time of the day. I was wondering a condition like “after no data movement for certain period”, but it doesn’t seem to work that way either.

Understanding the reset triggering conditions can help us better estimate impacts of specific FDB operations.

Thank you.

Leo

The value is tracked in memory on the data distributor, so it will reset any time the data distributor is re-recruited (i.e. due to some failures or the process restarting, etc). I don’t think it gets reset for any other reason.

Hi AJ,
So it’s an in-memory counter and increases over time. It only resets when the DD (or Master process as we run FDB v6.0.15) restarts or switches to another Stateless process.

I was testing for Stateless processes during last few days: restarting Stateless processes, or removing and adding Stateless nodes. That may explain the resets seen in the charter.

Thank you.

Leo