I'm not sure why it's happening though.
Bob, Bryan, Vince: can you comment on what might cause this? I understand that there is "slop" in the mean time between the scaler reads, but this seems pretty big. Perhaps a network stall? A glitch in the roll-over handling code (unlikely)? Could it be related to the 'really high single scaler read' glitch mentioned that's been noted since last night?
http://www.jlab.org/~adaq/halog/html/0902_archive/090219093347.html
A copy of this log entry has been emailed to: rom, vasulk, moffit
Figure 1