1. Please quit unused analysers. adaql5 has been running out of swap because a whack of idle analyser processes were sucking up 200MB of memory apiece.
2. Consider migrating your offline analysis to adaql6 instead of adaql5. adaql6 is also 25% faster than l5.
3. Don't run more scaler gui's than necessary. A couple of times I've
seen 4-5 copies of the src.tcl (BB and NA) scaler display running on
various workspaces of different machines. Similar restrictions should
apply to the xscalers program. There were some DAQ
problems in January traced to a whack of xscaler processes overloading
TS1.