[Halld-online] Online Meeting minutes

[David Lawrence]

Hi All,

  Sorry for the lateness in posting these minutes from last week’s Online Meeting …

https://halldweb.jlab.org/wiki/index.php/OWG_Meeting_21-Sep-2016 <https://halldweb.jlab.org/wiki/index.php/OWG_Meeting_21-Sep-2016>

Regards,
-David

Minutes[edit <https://halldweb.jlab.org/wiki/index.php?title=OWG_Meeting_21-Sep-2016&action=edit&section=5>]

Attendees: David L. (chair), Dave A., Graham H., Sean D., Naomi J., Carl T., Sergey F., Beni Z., Simon T., Curtis M., Bryan M.
Announcements[edit <https://halldweb.jlab.org/wiki/index.php?title=OWG_Meeting_21-Sep-2016&action=edit&section=6>]

ROOT6 is now the default version for hdops. Some performance issues reported. No show stoppers yet.
Collaboration Meeting[edit <https://halldweb.jlab.org/wiki/index.php?title=OWG_Meeting_21-Sep-2016&action=edit&section=7>]

Draft agenda is up
Online section moved to end of meeting to reflect current priorities of collaboration
Hovanes and Alex S. will be notified of their slots
DAQ[edit <https://halldweb.jlab.org/wiki/index.php?title=OWG_Meeting_21-Sep-2016&action=edit&section=8>]

Firmware testing still underway
Word on the street is that Alex S. has a working set of libraries in a private area but it is not yet installed in global area used by hdops
Street also believes all modules have had their firmware updated
Busy testing of fADC125 has not been done by Sergey since he is waiting on the DAQ system to become available
CODA 3.0.6 is ready, but has had only minimal testing due to the DAQ system being unavailable. Default is still CODA 3.0.5
Will do more CODA 3.0.6 testing when/if we have time prior to starting the FALL run. May need to do this during down times.
Monitoring[edit <https://halldweb.jlab.org/wiki/index.php?title=OWG_Meeting_21-Sep-2016&action=edit&section=9>]

Online monitoring system checkout underway
Some minor tweaks required with change to ROOT 6
Overnight run showed memory leak of ~2GB/hr
Will not be an issue for 2hr runs as planned, but may be if left on much longer (e.g. 6 hr)
Exhausting memory on farm nodes can cause them to stall in a way that requires a power cycle to recover
Will check on setting a ulimit to kill the process before this happens
RAID to silo data flow[edit <https://halldweb.jlab.org/wiki/index.php?title=OWG_Meeting_21-Sep-2016&action=edit&section=10>]

Work is ongoing but no significant changes to plan since last meeting

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-online/attachments/20160928/dcc7868b/attachment.html>