OWG Meeting 12-Aug-2015

From GlueXWiki
Jump to: navigation, search

Location and Time

Room: CC F326-327

Time: 1:30pm-2:30pm

Connection

You can connect using BlueJeans Video conferencing (ID: 120390084). (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
  2. Upgrade to RHEL7
  3. RCDB deployment
  4. DAQ system development
  5. AOT


Minutes

Attendees: David L. (chair), Simon T., Eric P., Curtis M., Dave A., Beni Z., Hovanes E., Sergey F., Bryan M., Alex S.

Announcements

  1. A directory /home/local has been created on gluon01-gluon05 that can be used to house local files. The motivating use case was the .mozilla directory used by Firefox. A safety feature built into Firefox prevented multiple computers from simultaneously using the same .mozilla directory. This should significantly reduce the number of times firefox refuses to open due to a profile already being in use. The downside is that bookmarks and profiles will not be propagated across all computers. They will need to be set for each individually.
  2. The firefox browsers have been set up to use the jprox proxy server so that websites outside of JLab may now be accessed.
  3. A Hall-D RHEL7 channel was set up some time ago for gluon01 and gluon24. This tagged a RHEL 7.0 version. The version installed on gluon46 was RHEL 7.1 and not tied to the channel so the plan is to wipe and load gluon46 so that it is aligned with the Hall-D channel. (See more below).
  4. A version of the LLVM/clang compiler has been built and installed on the gluons. It can be accessed via setup clang_3_6_2. Other packages available through the setup system can be listed by running "setup" with no arguments. Initial testing by Nathan Sparks indicates 30% less maximum RAM usage during compilation of sim-recon and >40% speed up in compilation time.
  5. Sergey and Dave believe they have solved the problem of "maximum number of clients reached". (See log entry)

Upgrade to RHEL7

  • We discussed upgrading all computers in the counting house to RHEL7 in September.
  • Hovanes noted that while the CSS GUIs have been tested on RHEL7, there has been no testing of other controls system software.
  • We clarified that computers that would be targeted for upgrade to RHEL7 would be:
    • EMUs (gluon40s and gluon50s)
    • Farm computers (gluon100s)
    • Desktops (gluon01-gluon05)
  • ROCs would continue to run 32bit RHEL5
  • Controls systems (gluon20s and gluon30s) could continue to run RHEL6 as these are not typically logged into by operators
    • The exception being gluon24 = gluon web which has already been upgraded to RHEL7 on the Hall-D channel
  • Sergey noted that the graphics on RHEL7 appears to be slower. Specifically, opening an new rcm.sh window, one can see all of the widgets being drawn. On the RHEL6 systems, the window draws too fast to see individual widgets being drawn.
    • n.b. though not mentioned at the meeting, we recalled later that gluon01 had a new video card installed when it was upgraded to RHEL7
  • Dave A. noted that there may need to be some optimization of the routing tables to get the best network performance after the upgrade to RHEL7

RCDB Deployment

  • Status has essentially not changed from last time
    • Dmitry identified some bugs in the RCDB web interface, but has been busy with other projects and unable to focus time on it. He hopes to do so soon.
    • Current DAQ system does fill in many fields whenever run is taken so database is getting filled
    • Sean's scripts have not yet been integrated, but that is still planned
    • Database replication is working so that the primary (gluondb1) is backed up to gluondb2 and also replicated to halldweb.

DAQ

  • Beni has tested an intermediate version of the fADC125 firmware.
    • This only implements a mode 7 type data.
    • Tests were able to handle 120kHz trigger rate with block level 20, buffer level 4
    • Thresholds set such that 1 hit per chip read out (where each chip handles 6 channels)
  • Dave A. reported that CODA 3.04 is ready for deployment.
    • He will install in Hall-D as he has done in the past for other CODA versions
    • Includes cMsg 4.0 which is not backward compatible with cMsg 3.X
    • Includes new ROC program that has cMsg interface for executing commands received from remote hosts
    • Several places where there is incompatibility with CODA 3.03 programs so switch to 3.04 must be complete (i.e. not partially done)