OWG Meeting 1-Jun-2016

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

• Previous Meeting (18-May-2016)

Agenda

1. Announcements
   • 3rd RAID server
   • New Online Farm nodes
   • gluon01: RHEL7-> RHEL6
   • Whiteboard cameras
2. Missing RunLog.tar files
3. gluon kernel versions
4. Compiler upgrade (gcc 4.9.2)
5. Write-through cache for raw data
   • First 5 raw data files will be automatically cached
6. DAQ
   • CODA 3.0.6
7. L3 Status
   • Level-3 Trigger Meetings
   • post-module real data playback
8. ROL status (SYNC events)
   • Scaler Data format
9. AOT

Minutes

Attendees: David L. (chair), Beni Z., Sean D., Simon T., Naomi J., Dave A., Carl T., Curtis M., Mark I.

Announcements

• Power Outage
  • There was a power outage on the accelerator site yesterday evening for about 1.5 hours starting at around 19:30
  • Techs were notified (not Physicists)
  • Gluon computers did not turn off
  • Helium in supply tank was lost (but not in magnet)
  • Gas system survived on UPS
• 3rd RAID server will be added to large IT procurement at time of award
• New Online Farm nodes:
  • Nature of them is still up in the air (i.e. exactly the type of CPU). Some benchmarking by IT people currently underway
• gluon01: RHEL7-> RHEL6 : still not done. David will remind Paul so it stays on his radar
• Whiteboard cameras: installed but still not accessible. Will need help from Hovanes

Missing RunLog.tar files

• Near end of Spring run noticed numerous RunLog.tar files were missing from tape
• Belief is that this is due to intereference between two cron scripts
• A fix has been implemented in the wrapper script run on the RAID servers that hopefully fixes this issue
• Won't really know until we go back into production data taking
• E-log entry made to document changes

Gluon kernel versions

• Gluons running RHEL6 have different kernel versions
• A "yum update will need to be done on each node to bring them up to date and in alignment with one another
• David will schedule this and send out and announcement since this will require a reboot of all computers

Compiler upgrade (gcc 4.9.2)

• Offline made gcc4.8 or higher a requirement as of today
• Proposal put forth to make some computers (ones used mainly for controls) default to gcc4.4.7 while the rest default to gcc 4.9.2 following a suggestion by Sergey at last meeting
• Hovanes not at meeting to discuss so David will speak with him offline

Write-through cache for raw data

• Suggestion was made just prior to Spring run to use write-through cache for transferring raw data to tape. This would leave most recent files on cache for offline monitoring
• We deferred implementing this until this summer to avoid potential issues with data taking in the Spring
• It turns out SciComp already has a system for moving files matching certain naming patterns from the staging disk to cache. They have now included an appropriate pattern for the first 5 raw data files from any future runs.
• Copying to tape via staging disk remains as it has been.

DAQ - CODA 3.0.6

• Dave A. has installed CODA 3.0.6 on the gluon cluster and it is ready for testing
• One key new feature is that it now provides an option from transferring data via direct socket connection rather than through ET
  • New jcedit has option to select for each component how it will receive its input
  • Dave A. noted that if the COOL config. were changed to use direct sockets and then one tried using that config with CODA 3.0.5, it would likely not work. We will need to remember to either change the configuration back, or keep a backup of the COOL config used with CODA 3.0.5
• JinFlux support is now also built into CODA 3.0.6

L3 Status

• Biweekly meetings have started
• David L. plans to resurrect system for playing back data from ROCs stored on a RAM disk
  • We can use real data this time instead of MC data
  • Beni will be providing a variable rate random trigger to the TS front panel for DAQ rate testing that can also be used for L3 testing.

ROL status (SYNC events)

• Alex S. (not at meeting) has committed some code to the repository for handling data from f250 modules in sync events
• Changes were made to current default parser, but will need to be ported to new parser (David will help with that)