Project & storage
Storage
Per RX day
Each column is one RX day. Split RXs sit side by side under a shared bracket; the days still to shoot carry the forecast.
Streams against its pool
Running total and projection for the Streams storage on its own scale.
PSC against its pool
Running total and projection for the PSC storage on its own scale.
The log
One row per RX day, shared by both modes. Figures are per mode; a day with none against it in this mode is one still to come, and those are what the forecast runs over.
| MAIN | PARITY | ||||||
|---|---|---|---|---|---|---|---|
| RX | Date | Streams TB | PSC TB | Cumulative | Streams TB | PSC TB | |
How the projection is made
The forecast takes the mean of the last few recorded RX days and carries it forward over the RX days that have no figures yet, projecting Streams and PSC separately against their own pools.
Main and Parity share one schedule and nothing else. The RX days, their splits and their dates are one list, so laying in the schedule once covers both. Media sizes and both storage pools are held per mode, because the parity set is its own copy on its own drives — and the RX days left differ too, since a day can be logged in one mode before the other.
The days the forecast runs over come from the log itself, not a number you keep up to date: any line item without figures in this mode is a day still to come. Set RX days in project to lay in the whole schedule at once, then fill each day in as it wraps.
The band is one standard deviation of a day’s figure, widened over the run as the square root of the days left — the spread of a total, not one day repeated. It does not forecast below three recorded days, and never runs past the days in the log. Split RXs count as the days they are: RX04A and RX04B are two RX days and one RX, so the per-day and per-RX figures differ, and the projection uses the per-day one.
Quoted against sent. Two figures per pool, because they are rarely the same number. Quoted is what the production is paying for; sent is what actually went out on the drives, which is usually more so there is contingency. RigSpace judges against the quoted figure first and treats the drives as the safety net behind it — the other way round, a meter reading comfortably full of spare capacity hides the fact that the contingency is being spent. Where the drives hold more than was quoted, the meter carries a marker at the quoted figure and both pool charts draw two lines: QUOTED solid, SENT dashed behind it. The RX day the projection crosses the quote is marked on the chart, because that is the day worth raising with a producer rather than at wrap.
When a pool reads as tight or over. Each pool is judged on its projected total — what is banked plus what the run rate says is still to come — against the figure you entered for it, and the two pools are judged separately in each mode. Under 85% of the quote it reports what is left and stays quiet. From 85% to 100% it reads tight: the forecast fits but has no useful margin, and one heavy day or one added RX will take it past. Over the quote it reads over and gives the overshoot in TB and per cent, saying whether it is still inside the drives you sent or past those too. With no quoted figure entered it falls back to judging against what was sent. Those bands describe the forecast, not what is on the drives — a pool can read tight while less than half full.
All figures are decimal TB (1 TB = 1000 GB).