OWG Meeting 18-May-2016

From GlueXWiki
Jump to: navigation, search

Location and Time

Room: CC F326

Time: 2:00pm-3:00pm

Connection

You can connect using BlueJeans Video conferencing (ID: 120 390 084). (Click "Expand" to the right for details -->):

(if problems, call phone in conference room: 757-269-6460)

  1. To join via Polycom room system go to the IP Address: 199.48.152.152 (bjn.vc) and enter the meeting ID: 120390084.
  2. To join via a Web Browser, go to the page https://bluejeans.com/120390084.
  3. To join via phone, use one of the following numbers and the Conference ID: 120390084
    • US or Canada: +1 408 740 7256 or
    • US or Canada: +1 888 240 2560
  4. More information on connecting to bluejeans is available.


Previous Meeting

Agenda

  1. Announcements
    • RAID disk upgrade (no movement)
    • Write-through cache for raw data (no movement)
    • Farm upgrade
      • 8 new nodes (~2x as fast as current 12 farm nodes)
    • gluon01: RHEL7-> RHEL6
    • Whiteboard cameras
  2. Compiler upgrade (gcc 4.9.2)
    • sim-recon will require it starting June 1st
    • testing required: does having newer glibc in path interfere with programs compiled using older glibc?
  3. JinFlux slides
  4. Reboot Notifications
  5. DAQ
  6. L3 Status
  7. ROL status (SYNC events)
    • Scaler Data format?
  8. AOT

Minutes

Attendees: David L. (chair), Curtis M., Sergey F., Carl T., Beni Z., Vardan G., Naomi J., Sean D., Dave A., Alex S.

  1. Announcements
    • Regarding new farm nodes: Dave A. noted that these will be very new CPUs and may not be fully supported by RHEL6. We may need to run RHEL7 on them. We'll address that once they arrive and the CC can explain the issues to us.
  2. Compiler upgrade (gcc 4.9.2)
    • No significant discussion here since Hovanes was not present. Sergey suggested we could consider specifying some computers as having the "controls" environment by default. This would have gcc 4.4.7 be the default compiler. All other computers could then default to a different compiler.
  3. JinFlux slides
    • Vardan presented some slides on the JinFlux package he authored which provides a Java interface to an Influx DB
    • The Influx DB is useful for information that has a limited shelf life before becoming obsolete
    • The Grafana package provides a way to build graphical displays of the data contained in the DB
    • Vardan has both the DB server and the Grafana server running on claraweb.
    • We could write values into it from our DAQ system and then build a dashboard that could display it offsite
    • He showed a working example at http://claraweb.jlab.org:3000
  4. Reboot Notifications
    • David has implemented a system to send e-mail notification to him whenever a node reboots
    • This may help identify which nodes are problematic and how often spontaneous reboots are occurring
    • It was noticed that some nodes seem to reboot at the same time every day or every couple of days
    • No one else expressed interest in getting these e-mails (??)
  5. DAQ
    • Several new firmware versions are now (or very soon will be) available
    • These include fADC125, fADC250, CTP, SSP, F1TDC
    • Cody and Ben were going to do a single crate test using the BUSY mechanism and Sergey will help with a test that includes the TS
    • Alex S. will be installing and testing the other firmware over the summer starting sometime in June
      • He will start with the new fADC250 firmware which has many new features and provides data in a new format
      • Alex will provide David with an example file and a format specification once the initial testing is done and they become available
      • Playback features in the new firmware will allow tuning of the trigger parameters
      • Different holdoffs may be programmed for different triggers allowing the first few triggers to have a smaller holdoff than the last which needs a longer holdoff
      • Up to now, we have used only a single holdoff for all triggers. Alex will do some measurments this summer to tune these in order to maximize rate/livetime
    • CODA 3.06 is nearly ready for release. Dave A. expects to deploy it by the end of next week
    • Sergey noted that once we get past the current bottleneck due to the holdoff time then we may quickly encounter another bottleneck due to the DMA setup/transfer time
      • The argument was that there is a 4MB limit on a single DMA transfer. This limits the number of events that can fit inside a block when each fADC125 has full occupancy
      • There was some discussion on this but details would need to be hashed out offline
    • Sergey noted that there is also a potential bottleneck looming from the tagger crate due to it reading out 60-70MB/s on a 1Gb interface
      • Dave A. said he has a couple of 10Gb cards we could test
      • Alex S. noted that the tagger readout window is quite large and could be reduced which would have a significant impact
  6. L3 Status
    • Meeting tomorrow at 3pm
  7. ROL status (SYNC events)
    • Some data has been implemented and Alex has plans to include more diagnostics information
    • Documentation doesn't yet exist, but he will supply code or a document once things are more final
    • Sync events in Spring data are currently ignored by the sim-recon parser