Closed strilov closed 3 years ago
There are a lot of warning messages being produced by the new code (examples below).
Could it be that the test should be against 1000000 (one million microseconds)?
2021-May-27 11:42:47,227 WARNING [void dunedaq::timinglibs::TimestampEstimator::estimator_thread_fn(std::unique_ptr<dunedaq::appfwk::DAQSource<dunedaq::dfmessages::TimeSync> >&) at /home/biery/dunedaq/27May/sourcecode/timinglibs/src/TimestampEstimator.cpp:87] The most recent TimeSync message is behind current system time by 68745 us.
2021-May-27 11:42:47,237 WARNING [void dunedaq::timinglibs::TimestampEstimator::estimator_thread_fn(std::unique_ptr<dunedaq::appfwk::DAQSource<dunedaq::dfmessages::TimeSync> >&) at /home/biery/dunedaq/27May/sourcecode/timinglibs/src/TimestampEstimator.cpp:87] The most recent TimeSync message is behind current system time by 78863 us.
2021-May-27 11:42:47,247 WARNING [void dunedaq::timinglibs::TimestampEstimator::estimator_thread_fn(std::unique_ptr<dunedaq::appfwk::DAQSource<dunedaq::dfmessages::TimeSync> >&) at /home/biery/dunedaq/27May/sourcecode/timinglibs/src/TimestampEstimator.cpp:87] The most recent TimeSync message is behind current system time by 88990 us.
Thanks Kurt. Fix pushed.
Looks good. For future tests the clocks at np04 just have to be synchronized a bit better.
Using Florian's config, where the fake
HSI
emulator is running onnp04-srv-025
, and the readout apps running onnp04-srv-028
, I can indeed see TimeSync messages with system time around 90s behind current system time. This does seem to correspond to the different in date/time seen betweennp04-srv-025
andnp04-srv-028
, where025
is ahead of028
.Is there anything else we should do in the code?
Can we run NTP at NP04, especially for these types of tests?