OWG Meeting 9-May-2012

From GlueXWiki
Jump to: navigation, search

Agenda

  • Announcements
    • Summer students and projects
  • Review of minutes from 25-Apr-2012 meeting
  • Network monitoring at JLab - Bryan Hess
  • Agenda for the Collaboration Meeting May 21-23
  • EVIO status report - Carl, Elliott
  • Recent online plans and purchases - Hovanes
  • Barcode scanner - Elliott
  • Hall D Component Inventory Database - Elliott
    • telecom info, location hierarchy
  • Counting House status - Hovanes
  • 12GeV Schedule - Hovanes, David, Elliott
  • Elog and Accelerator division projects - Graham
  • DAQ group - Bryan, Dave A, Graham
    • CODA3, EVIO, ECS/CSS, raw event file, disentangler, readout list
  • EPICS - Hovanes, Vanik
    • transferring some projects to Accelerator Div.
  • Controls - Elliott
  • Electronics, trigger and DAQ status reports - Chris, Alex, Dave, Fernando, Ben, Ed, William, Bryan
  • Mantis Task Tracker
  • Midas Elog


  • Calibration/Alignment issues?
  • Computer system manager


Time/Location

1:30 PM Wed 9-May-2012 CC F326


Announcements

Let Elliott know if you have projects for summer students. So far we've only attracted one student from CNU, and we could use more.


Next Meeting

Note: Collaboration Meeting 23-May-2012

1:30 PM Wed 6-Jun-2012 CC F326


Minutes

Present: Elliott W, Simon T, Bryan Hess, Carl T, Mark I, Bryan M, Vanik K, Beni Z.


Network Monitoring

Bryan Hess, head of networking in the computer center, explained how they monitor hundreds of networking switches at the lab, see his slides for many details.

  • All monitoring done in-band.
  • Extensive checking and archiving of results for later analysis.
  • Monitor latency, packet loss, throughput as early warning of future switch failure.
  • Network configurations stored in svn.
  • They do NOT monitor "edge" devices (e.g. computers, about 5000 total), just the networking infrastructure.
  • They also do not monitor end-user applications, services, etc.
  • Monitoring system is "JNet", written by Bryan and Brent.
  • Switches indexed by location, gives information about what is connected to the switch, including phones.
  • Nightly and continuous checking.
  • 24x7 support for networks by CC (via calling Guard Shack), Bryan and/or Brent gets called as needed.
  • Accelerator manages completely separate networks, almost nothing in common with CC networks.
  • Best way to reset switch is to cycle power, thus we need an out-of-band power reset system.
  • Bryan is happy to monitor our network equipment in the JNet system.

Finally, Bryan says it should be no problem feeding alarms from JNet into the Hall D online alarm system.


EVIO Status

Carl and Elliott gave a report on the latest changes to EVIO, now called EVIO version 4. Most of them have been driven by CLAS requirements, but the improvements will benefit Hall D as well. V4 is backwards-compatible with previous versions, but older versions cannot read V4 files.

  • Carl has append mode working, needs to be checked in.
  • He is working on random-access I/O.
  • Fixed-length blocks have been eliminated, now they are variable length depending on the number of events or bytes.
  • I/O has been unified under a "stream" model...file I/O and the new buffer and socket I/O all work the same way.
  • Banks and segments keep track of the number of padding bytes added to arrays of shorts or bytes to reach the 4-byte boundary.
  • Composite banks have been added that implement packed row-wise data storage. The idea and swapper came from CLAS (Sergey B).
  • Name to tag/num dictionaries are now handled as part of the API. Dictionaries are just XML strings, and are automatically stored in the file or buffer.
  • A no-copy read facility was added for speed.
  • Mechanisms to get fast access to banks in the underlying serialized buffer was added, no need to deserialize into the full object tree if you just want a few banks out of many.


Purchases

  • Hovanes has been talking about purchasing a terminal server. Bryan M noted that you still need to make an RS232 connection to the ROC's to configure them, but the connection won't be used as much in Linux as it was in VxWorks.
  • Elliott is investigating barcode scanner systems to tie into the JInventory system.


FADC125

Initial testing looks good. Beni needs about 36 modules for the full FDC test.


Misc

  • The first summer student arrives Monday. Andrew will work on two-way JANA to DED (event display) communication.
  • Elliott added many new tasks to the Mantis, many suitable for summer students.
  • So far we only have one summer student.
  • Elliott is continuing to improve the JInventory system. Sergey will add a cable component. Eventually we need web pages suitable for hand-held devices.
  • The last cabinet for the counting house UPS should be installed next week.
  • Still don't know what to do about long-lifetime FCC Class B UPS power in the hall.
  • Elliott, Dave and Hovanes continue to work on 12GeV project lines and reports.
  • Vanik continues to work on stepper motor control
  • Accelerator will completely take over 4 Hall D power supplies, leaving us with just the solenoid supply. And they will help with routine maintenance and repairs for that one as well.
  • Accelerator will thus locate VME crates in the hall and tagger area, and will run their own networks to them.
Retrieved from "https://halldweb1.jlab.org/wiki/index.php?title=OWG_Meeting_9-May-2012&oldid=65649"