Skip to content

Problems running the latest replay on the farm

Hi Sylvester,

There seems to be a problem with current master branch in hallc_replay: https://eicweb.phy.anl.gov/jlab/hallc/exp/polhe3/hallc_replay.git

After syncing with upstream I get similar errors to Junhao. I'm guessing that the new hallc_replay/master requires an updated singularity container that needs to be rolled out?

For now Junhao will jump back to commit 9f79255c and run tests from there. Should be good for the moment.

-- Brad

On Wed, 27 Jan 2021, Junhao Chen wrote: Hi Brad, I encountered a problem running hcreply. And I get the same error in both my working repository and a brand new repository forked from the gitlab repository. The error messages shows: "

Processing
SCRIPTS/src/replay_shms.cxx+(11498,5000,0,"all","DEF-files/SHMS/PRODUCTION/pstackana_production_all.def","DEF-files/SHMS/PRODUCTION/CUTS/pstackana_production_cuts.def")...

In file included from replay_shms_cxx_ACLiC_dict dictionary payload:8:
In file included from ././SCRIPTS/src/replay_shms.cxx:8:
In file included from /usr/local/include/spdlog/spdlog.h:12:
In file included from /usr/local/include/spdlog/common.h:36:
In file included from /usr/local/include/spdlog/fmt/fmt.h:23:
/usr/local/include/spdlog/fmt/bundled/format.h:208:32: error:
non-constexpr declaration of 'const_check' follows constexpr

declaration template inline T const_check(T value) { return value; } ... ...

" I also tried to let Mingyu help verify. While his repository works, the one I copied from his repository do not work. This time a different error message shows: "

[51768] [run] [info] Overwriting existing output file:
ROOTfiles/shms_replay_production_all_11498_5000.root
terminate called after throwing an instance of 'std::runtime_error'
  what():  <THcParmList::LoadParmValues>: Could not find `gpbeam` in
database!
/group/c-polhe3/software/opt/hallac_container_1.8/bin/hcana: line 20:
51739 Aborted
/apps/singularity/singularity-3.5.2/bin/singularity exec -B /site:/site
/group/c-polhe3/software/opt/hallac_container_1.8/lib/hallac-1.8.0.sif
hcana $@
(reverse-i-search)`o': ln -s
/volatile/hallc/c-polhe3/mychen/example_^Ctput/ ROOTfiles

". I linked raw and ROOTfiles directory to /cache/hallc/c-polhe3/raw and /volatile/hallc/c-polhe3/jhchen/output/. I sourced the file /group/c-polhe3/software/setup.sh. Do you have any idea what is wrong? Thank you, Junhao

On Jan 27, 2021, at 1:10 PM, Brad Sawatzky <[2]brads@jlab.org> wrote: Hi Junhao,

The container is working for me on the ifarm.

  • Be sure you are not running any 'extra' ROOT or hcana setup files in your shell init scripts.

  • Do not source any other root-setup scripts

  • Do not source any other /site/12gev_phys/softenv.* scripts

  • Be sure you are running the 'hcana' binary and not 'root' to run your scripts.

This is all you should need to do to set up the correct environment: % source /group/c-polhe3/software/setup.sh ## If your shell is bash -- OR -- % source /group/c-polhe3/software/setup.csh ## If your shell is tcsh

Then cd into your replay directory % cd /group/c-polhe3/Users/$USER/hallc_replay/ % which hcana /group/c-polhe3/software/opt/hallac_container_1.8/bin/hcana % hcana hcana [0] .x SCRIPTS/src/replay_shms.cxx+(11498, .... )

-- Brad

Hi Brad,

I only source /group/c-polhe3/software/setup.sh. And I do use bash. I checked the shell initialisation scripts: .cshrc, .zshrc and .logon.

  • The only thing I added is 'exec bash' in the end of .logon
  • There is a line 'source /site/env/syscshrc' in .cshrc, I am not sure what is this. I did replay step by step:
  • source, cd into hallc_replay
  • confirmed hcana is /group/c-polhe3/software/opt/hallac_container_1.8/bin/hcana
  • run hcana, then in hcana run the '.x SCRIPTS/src/replay_shms.cxx+(11498, .... )'
  • it still do not work with the same error Then I check like this:
  • in hcana run a single line '#include "spdlog/spdlog.h"',
  • the same error message comes up as if I run '.x SCRIPTS/src/replay_shms.cxx+' in hcana, or run `hcreplay ...' command lines
  • scroll back the hallc_replay repository to commit `9f79255c' where the recent '[1]Resolve "Issue with hard-coded root build-dir"' commits have not been applied.
  • this time I have no problem running "spdlog/spdlog.h" in hcana.
  • but if I do a replay, I will encounter another error with this scrolled back repository.
  • this time error message is
[220882] [run] [info] Creating new output file:
ROOTfiles/shms_replay_production_all_11498_5000.root
THaCodaFile: ERROR while trying to open
./cache/shms_all_11498.dat: No such file or directory

Also there are folders .singularity and .root_build_dir in my home directory. I did not change them. The latest update time for them are both "Apr 15 2020". Do you see any clues? Thank you, Junhao

@Sawatzky