Abstract
When I prepared various things for tomorrow's maintenance about the vacuum system update (see also klog#32511), I noticed pressure values around EXA cannot be read.I restarted a runaway process and fixed this issue.
Now status of vacuum DAQ should be normal.
Details
Precisely, pressure value could be read but connection to RaspPi@EXA took roughly 10min. per once and execution of old CC-10 script (vacuum.sh) also took roughtly 10min. For this reason, recent DAQ for all vacuum channels seemed to be executed as ~20min cadence without any notification. (I'm not sure when this issue started.)At first, I logged-in to RaspPi@EXA and searched problematic process. Then I found soon load average was over 7.0. But there was no process which used lots of CPU though CPU usage by swap process was slightly high (~20%), there was no process. Because swap process worked hard, I checked RAM situation and found free RAM was only ~10MB. And also "lxpanel" process used ~90% of RAM, which was related to the GUI (This is the same trouble as the hang-up of workstation in the control room). For the stable operation, muti-user target might be better than the graphical target. Anyway, I restarted lxpanel process and then problem was solved.