[Jamie, EricQ, Jenne, Diego]
This is something that we discussed late Friday afternoon, but none of us remembered to elog.
We have been noticing that EPICS seems to run pretty slowly, and in fact twice last week froze for ~2 minutes or so (elog 10756).
On Friday, we plotted several traces on StripTool, such as C1:SUS-ETMY_QPD_SUM_OUTPUT and C1:SUS-ETMY_TRY_OUTPUT and C1:LSC-TRY_OUTPUT to see if channels with (supposedly) the same signal were seeing the same sample-holding. They were. The issue seems to be pretty wide spread, over all of the fast front ends. However, if we look at EPICS channels provided by the old analog computers, they do not seem to have this issue.
So, Jamie posits that perhaps we have a network switch somewhere that is connected to all of the fast front end computers, but not the old slow machines, and that this switch is starting to fail.
My understanding is that the boys are on top of this, and are going to figure it out and fix it.