Project

General

Profile

Bug #1176

error message in DAQ unpacking

Added by Rogers, Chris almost 11 years ago. Updated almost 11 years ago.

Status:
Open
Priority:
Normal
Category:
RealData
Target version:
Start date:
23 November 2012
Due date:
% Done:

0%

Estimated time:
Workflow:
New Issue

Description

I get an error message while unpacking DAQ. I think this is causing the reconstruction to fail.

I got triggers in spills 0, 2, 6, 12 ...
I got unpacking errors in maus event 2, 8, 20, 38 ... (== n*3+2)

The error message is:

ERROR in MDdataContainer::Get32bWordPtr - the size is exceeded (i=11020 size=48).
*** Unpacking exception in MDfragment::InitPartEventVector
while processing V1731 fragment (board Id: 9)
*** Unpacking exception in MDprocessManager::ProcessSubEvent :
while initializing the fragment
*** Unpacking exception in InputCppDAQData::getCurEvent() : 
DAQ Event skipped!

I attach full log for run 4693 and MAUS execution script (command line used to execute is somewhere near the top of the log).


Files

analyze_data_online_single_threaded.log (53.4 KB) analyze_data_online_single_threaded.log Rogers, Chris, 23 November 2012 14:55
analyze_data_online_input_transform.py (2.8 KB) analyze_data_online_input_transform.py Rogers, Chris, 23 November 2012 14:55
04693.tar (130 KB) 04693.tar Rogers, Chris, 30 November 2012 11:32
04693.tar.md5 (44 Bytes) 04693.tar.md5 Rogers, Chris, 30 November 2012 11:32
#1

Updated by Rogers, Chris almost 11 years ago

Nb: we arent getting any data from reconstruction, maybe this is the cause...

#2

Updated by Karadzhov, Yordan almost 11 years ago

Seems like you use the old unpacking that has no support for the Zero Suppressed data.
Call me on my mobile phone.
Yordan

#3

Updated by Karadzhov, Yordan almost 11 years ago

Or get the code from launchpad (bzr branch lp:unpacking).
Y.

#4

Updated by Rogers, Chris almost 11 years ago

I added the command line tag

-Enable_V1731_Unpacking=False

to disable V1731 (Ckov flash adc unpacking?) and indeed now I can TOF reconstruction to slab hits in TOF1, which is all we expect with cosmics.

Seeing Yordan comment above, I will try to get the latest version of unpacking.

#5

Updated by Rogers, Chris almost 11 years ago

I reran the unpacking install script at third_party/bash/53unpacking.bash but the reconstruction still fails as described above. I thought we agreed that this was the way to install unpacking. Can you please supply installation instructions and some release management for your code?

#6

Updated by Karadzhov, Yordan almost 11 years ago

installation instructions:
1. get the code (bzr branch lp:unpacking)
2. copy *.h and *.cpp in maus/third_party/build/unpacking-maus/
3. cd maus/third_party/build/unpacking-maus/; scons
4. cd ../../../; scons

#7

Updated by Rogers, Chris almost 11 years ago

I do

bzr branch lp:unpacking

from either my laptop or the onrec01 and get
bzr: ERROR: Not a branch: "bzr+ssh://bazaar.launchpad.net/%2Bbranch/unpacking/".

#8

Updated by Rogers, Chris almost 11 years ago

Okay, with some help from Yordan on the phone I managed to build unpacking... but the error message is still appearing. I did manage, briefly, to get some TOF slab hits through by setting Enable_V1731_Unpacking=False but for some reason when I tried rerunning with this flag set I still get the same error message (in V1731 still) - so maybe something else going on?

#9

Updated by Rogers, Chris almost 11 years ago

I attach the tarball 04693.tar plus md5 (with thanks to Henry).

#10

Updated by Rogers, Chris almost 11 years ago

Just had a phone call with Yordan regarding unpacking.

Firstly, Yordan has fixed the bug that was causing MAUS online recon to crash. There were two issues, one in the unpacking library and another in InputCppDAQData (the MAUS code that calls unpacking). A new version of unpacking has been installed in the control room branch on onrec01 PC. I will update the issue tracker 1176 to this effect.

We should test on cosmics, can we do that tomorrow morning please Yagmur? Yordan I forgot to ask, will you be around?

We also had a conversation about the unpacking release management/MAUS install. Agreed:
  • we will release unpacking code as a tarball.
  • Yordan will put it on a reasonably reliable webserver, to be decided where, probably uni geneve hosted.
  • version name will be in the tarball name
  • Yordan will email some relevant mailing list, probably mice-software, when a new version is released.
  • Rogers will manage a script that does a wget/cmake/make in MAUS third party installers.
  • Rogers will add dependency on cmake to MAUS external dependencies (yum -i cmake, this is also a geant4.9.5 dependency so will be needed sometime anyway)

Also available in: Atom PDF