jgarofoli/LCLS-data-summary

test on other hutches

Closed this issue · 6 comments

test on other hutches

mpirun -n 6 python launcher.py AMO/amoc8114 104 -M 400

test on AMO, got this error msg:

2015-03-03 11:38:24.138 [FTL] {PSXtcInput::Index} Index.cpp:282 - Index mode does not support multiple epics archiving intervals. Variable 0 interval: 0.005, Variable 86 interval 0.008
 This error can be turned into a warning by setting the parameter allow-corrupt-epics in the [psana] section of the configuration file, but slow EPICS data will be corrupt.
[psana1201:10001] *** Process received signal ***
[psana1201:10001] Signal: Aborted (6)
[psana1201:10001] Signal code:  (-6)
2015-03-03 11:38:24.138 [FTL] {PSXtcInput::Index} Index.cpp:282 - Index mode does not support multiple epics archiving intervals. Variable 0 interval: 0.005, Variable 86 interval 0.008
 This error can be turned into a warning by setting the parameter allow-corrupt-epics in the [psana] section of the configuration file, but slow EPICS data will be corrupt.
[psana1201:10003] *** Process received signal ***
[psana1201:10003] Signal: Aborted (6)
[psana1201:10003] Signal code:  (-6)
2015-03-03 11:38:24.138 [FTL] {PSXtcInput::Index} Index.cpp:282 - Index mode does not support multiple epics archiving intervals. Variable 0 interval: 0.005, Variable 86 interval 0.008
 This error can be turned into a warning by setting the parameter allow-corrupt-epics in the [psana] section of the configuration file, but slow EPICS data will be corrupt.

SXR tested successfully.

Similar problem on xcs:

2015-03-03 11:43:52.358 [FTL] {PSXtcInput::Index} Index.cpp:282 - Index mode does not support multiple epics archiving intervals. Variable 0 interval: 1, Variable 162 interval 0.1
 This error can be turned into a warning by setting the parameter allow-corrupt-epics in the [psana] section of the configuration file, but slow EPICS data will be corrupt.
[psana1201:10378] *** Process received signal ***
[psana1201:10378] Signal: Aborted (6)
[psana1201:10378] Signal code:  (-6)

AMO works on a newer run.

MEC works, but will fail with parallel processing due to so few events. So, the regularizing of final jobs needs to be addressed soon.

All hutches work.

Issues (to be addressed in separate tickets):

  • not enough events will cause the parallel jobs to hang, so this needs to be addressed.
  • Off-by-one should report that it has not found enough data, rather than "False".