next up previous contents index
Next: astrod and rotsed Up: Rotse-III Scheduling Previous: Rotse-III Scheduling   Contents   Index


Introduction

The ROTSE-III scheduler is a modified queue scheduler implemented in astrod, described in detail in Section 1.2.8. Although the priority of a specific schedule item can be set (this is essential for fast alert response), astrod typically uses a scoring algorithm to decide which schedule item should be carried out next.

There are three levels of scheduling: guest user observations, prompt burst responses, and the regular telescope operations such as sky patrols. The instructions for these operations are stored in three files: astrod_rtrig.conf, astrod_rsched.conf, and astrod_lsched.conf. These files should only be altered directly by a qualified ROTSE technician. For guest observers, there is a Perl script, called user_sched.pl (See Section 4.6) that provides an interface to implement and manage their requested observations.

The standard ROTSE operational instructions consist of ``trigger'' definitions, which define a sequence of images to be taken, and ``schedule'' definitions, which provide a set of parameters to astrod to let it decide when and how to activate the trigger protocols. Standard triggers include burst responses, dark runs, runs to construct flats, and focus runs, as well as a ``sky patrol'' on a regular grid on a specified region of sky. Guest users may define triggers and schedules to carry out observing programs on specific desired targets. Each trigger has an associated three letter acronym (TLA) that becomes the root of the image filename for easy recognition and sorting later. How to set these various configuration parameters is described in this chapter.

The schedule is read in from astrod.conf on system startup (when rotsed is first launched), and near the start of each night's observing (when the sun dips below 5$^\circ$ above the horizon). In addition, although you should never do this under normal circumstances, sending a SIG_HUP signal to astrod erases the current schedule and re-reads the configuration file. This can be accomplished with the following (the process id is only an example):

$ ps auxww | grep astrod

root 20345 ... /rotse/run/bin/astrod ...

$ kill -HUP 20345

I should write a little perl script that does this automatically.


next up previous contents index
Next: astrod and rotsed Up: Rotse-III Scheduling Previous: Rotse-III Scheduling   Contents   Index
Rotse Pager 2003-05-20