Troubleshooting the Trigger System

The first thing you should do is "reboot trigger". If this does not work, look at "show component" from the RUN GUI on RTS02. All Trigger components/crates should have a white background. Look for any component whose cell has a red background. These components are either turned off or not communicating with Run Control. All components should have an icon before their name. For components with a white background the icon should be blue. If the icon is a circle, they are included in the run. If the icon is a dash, they are not included in the run. To add or remove a component to the run, simply use the mouse and click on the icon before starting a run. The icon on any system with a red background must be a dash or the run will not start. The only components you MUST have in a run are L1, L0, L2 (and TCD if other systems (e.g. TPC, BSMD, etc.) will be included in the run).

To reestablish the connection to run control you might try rebooting the offending CPU from its window in the Trigger desktop (startrg.starp.bnl.gov). Each trigger CPU has a separate window on this desktop, usually iconized for space. If you don't see a window or an icon for the CPU in question, go to the terminal window on the desktop and type "grab xxx" where xxx is the name of the component given on the “show component” GUI on the Run Control machine, RTS02. For instance, if BBC is black, and you can't find the window, type "grab bbc" (without the double-quote marks). This should pop a new window on the desktop. Note: this prescription does not work for L2 as discussed in the next paragraph. The connection to these windows is maintained even when the connection to the run control is broken. To reboot from this window, just type "ctrl x" in the CPU's window. If this does not work, try power cycling the offending component via Slow Controls (i.e. go to the console behind the shift leaders desk labeled "sc5.starp.bnl.gov". look for "Medm" on the tool bar at the bottom of the screen. click on the tool bar and choose "vme_1st_plat.adl". if that option is not in the pull down list, then get in a terminal window and launch "vme_plat1". a new GUI will appear. find the icon with the name of the crate that you need to power cycle. click on the small pink/purple square in that icon and choose the first line of the pull down menu. a new GUI will appear. in that GUI click the "off" button. you should see the gauges and fan speed drop to zero. now click the "on" button. you should see the gauge and fan speed come back up. if not, try the "on" button again.) It is often the case that cycling power for the affected component will cure the problem and turn the icon blue. See "power cycling the trigger" in the operations/trigger area.

if you power cycle any crate, you need to restart L2 to get the STP connections to reset.

If L2 "hangs" or goes black, go to the trigger desktop, startrg.starp.bnl.gov, and find the TRG desktop. There you should find a window named "l2ana01". Cntrl-C in that window to restart the l2 processes or, if you do not find an "l2ana01" window or its icon on the desktop, generate one by clicking the "lightning bolt" icon on the tool bar.

The following crates are used in the trigger:

BBC and its digitizer crate BBQ (includes VPD & ZDC info) for vertex triggering

MIX and its digitizer crate MXQ for triggering on TOF, MTD and/or PP2PP

BCE, BCW, BC1 - for trigger on the Barrel and./or EndCap calorimeters

EPQ for triggering on the east EPD detector

FMS - with digitizer crates QT 1,2,3,4 for forward calorimeter on the west wall

SCALER48 for the scaler data

L1, L0, and TCD - these are the clock, L0 logic, and distribution crates

While the scalers are not required for any STAR running, it is essential for interpreting spin data. It is used for diagnostics during the HI running. If the scaler process hangs, and the above recovery procedures do not work, just remove the SCALER48 from the runs and contact the trigger group via email to startrg-hn.

Last updated 17-Jan-2017 -je