run number from env / to log

Author: sy-c

doc/README.md

@@ -382,6 +382,14 @@ In all replay modes, ReadoutEquipmentPlayer does not support LZ4 files or files
 Readout can cope with input files containing data from multiple CRUs. The data pages will be split and tagged accordingly to the RDH fields (respecting the "1 single link per page" CRU specification).
 
 
+### Bookkeeping
+
+Readout can store runtime statistics in O2 Bookeeping. See the readout.logbook* configuration variables.
+At runtime, the run number is provided by the O2 Control System when executing the START command.
+When running from the command line, a run number can be set using the O2_RUN environment variable.
+In interactive mode, the value is automatically incremented with successive START/STOP sequences.
+
+
 ### Frequently asked questions
 
 The [howto guide](howto.md) documents some typical use cases and will be extended according to users feedback.

doc/releaseNotes.md

@@ -570,3 +570,6 @@ This file describes the main feature changes for each readout.exe released versi
 - Added o2-readout-monitor-memory: to view in real time state of data pages.
 - Added consistency check of orbit vs timestamp when large gas in TF ids detected.
 - Updated configuration parameters documentation.
+
+## next version
+- When running from the command line, the environment variable O2_RUN can be used to set the run number. It is set to 0 by default, i.e. undefined run number.

src/mainReadout.cxx

@@ -142,7 +142,7 @@ static void signalHandler(int signalId)
 
 // some globals needed in other components
 std::string occRole;     // OCC role name
-tRunNumber occRunNumber = 0; // OCC run number
+tRunNumber occRunNumber = undefinedRunNumber; // OCC run number
 
 // a general purpose log function for DB
 void dbLog(const std::string &msg) {
@@ -228,24 +228,27 @@ class LogbookThread {
       // thread loop, 10Hz
       while (!shutdownRequest && (logbookHandle != nullptr)) {
         if (publishRequest.load() == 1) {
-          bool isOk = false;
-          try {
-            // interface: https://github.com/AliceO2Group/Bookkeeping/tree/main/cxx-client/include/BookkeepingApi
-            logbookHandle->flp()->updateReadoutCountersByFlpNameAndRunNumber(
-              occRole, occRunNumber,
-              (int64_t)gReadoutStats.counters.numberOfSubtimeframes, (int64_t)gReadoutStats.counters.bytesReadout, (int64_t)gReadoutStats.counters.bytesRecorded, (int64_t)gReadoutStats.counters.bytesFairMQ
-            );
-            isOk = true;
-          } catch (const std::exception& ex) {
-            theLog.log(LogErrorDevel_(3210), "Failed to update logbook: %s", ex.what());
-          } catch (...) {
-            theLog.log(LogErrorDevel_(3210), "Failed to update logbook: unknown exception");
-          }
-          if (!isOk) {
-            // closing logbook immediately
-            logbookHandle = nullptr;
-            theLog.log(LogErrorSupport_(3210), "Logbook now disabled");
-            break;
+          // publishing to logbook makes sense only if a run number defined
+          if (occRunNumber != undefinedRunNumber) {
+            bool isOk = false;
+            try {
+              // interface: https://github.com/AliceO2Group/Bookkeeping/tree/main/cxx-client/include/BookkeepingApi
+              logbookHandle->flp()->updateReadoutCountersByFlpNameAndRunNumber(
+                occRole, occRunNumber,
+                (int64_t)gReadoutStats.counters.numberOfSubtimeframes, (int64_t)gReadoutStats.counters.bytesReadout, (int64_t)gReadoutStats.counters.bytesRecorded, (int64_t)gReadoutStats.counters.bytesFairMQ
+              );
+              isOk = true;
+            } catch (const std::exception& ex) {
+              theLog.log(LogErrorDevel_(3210), "Failed to update logbook: %s", ex.what());
+            } catch (...) {
+              theLog.log(LogErrorDevel_(3210), "Failed to update logbook: unknown exception");
+            }
+            if (!isOk) {
+              // closing logbook immediately
+              logbookHandle = nullptr;
+              theLog.log(LogErrorSupport_(3210), "Logbook now disabled");
+              break;
+            }
           }
           publishRequest = 0;
         }
@@ -1357,6 +1360,17 @@ int Readout::_configure(const boost::property_tree::ptree& properties)
 
 int Readout::_start()
 {
+  // set run number for logs
+  theLogContext.setField(InfoLoggerContext::FieldName::Run, std::to_string(occRunNumber)); // this works also for undefinedRunNumber, 0 -> empty field in log API
+  theLog.setContext(theLogContext);
+  if (occRunNumber != undefinedRunNumber) {
+    setenv(envRunNumber, std::to_string(occRunNumber).c_str(), 1);
+    theLog.log(LogInfoDevel, "Run number %d", (int)occRunNumber);
+  } else {
+    unsetenv(envRunNumber);
+    theLog.log(LogInfoDevel, "Run number not defined");
+  }
+
   theLog.resetMessageCount();
   theLog.log(LogInfoSupport_(3005), "Readout executing START");
   gReadoutStats.reset(1);
@@ -1706,6 +1720,9 @@ int Readout::_stop()
 
 int Readout::_reset()
 {
+  // reset run number for logs
+  theLogContext.setField(InfoLoggerContext::FieldName::Run, "");
+  theLog.setContext(theLogContext);
 
   theLog.log(LogInfoSupport_(3005), "Readout executing RESET");
   gReadoutStats.counters.state = stringToUint64("> reset");
@@ -1914,15 +1931,6 @@ class ReadoutOCCStateMachine : public RuntimeControlledObject
     }
     // set run number
     occRunNumber = this->getRunNumber();
-    theLogContext.setField(InfoLoggerContext::FieldName::Run, std::to_string(occRunNumber));
-    theLog.setContext(theLogContext);
-    if (occRunNumber != 0) {
-      setenv(envRunNumber, std::to_string(occRunNumber).c_str(), 1);
-      theLog.log(LogInfoDevel, "Run number %d", (int)occRunNumber);
-    } else {
-      unsetenv(envRunNumber);
-      theLog.log(LogInfoDevel, "Run number not defined");
-    }
     return theReadout->start();
   }
 
@@ -1987,6 +1995,17 @@ class ReadoutOCCStateMachine : public RuntimeControlledObject
 };
 #endif
 
+// increment run number based on initial settings from env
+void incrementRunNumber() {
+  if (occRunNumber != undefinedRunNumber) {
+    occRunNumber++;
+  } else {
+    if (getenv(envRunNumber) != nullptr) {
+      occRunNumber = atoi(getenv(envRunNumber));
+    }
+  }
+}
+
 // the main program loop
 int main(int argc, char* argv[])
 {
@@ -2151,7 +2170,7 @@ int main(int argc, char* argv[])
         }
       } else if (theState == States::Configured) {
         if (theCommand == Commands::Start) {
-          occRunNumber++;
+          incrementRunNumber();
           err = theReadout->start();
           if (err) {
             newState = States::Error;
@@ -2234,7 +2253,7 @@ int main(int argc, char* argv[])
       return err;
     }
 
-    int nloop = 1; // number of start/stop loop to execute
+    int nloop = 3; // number of start/stop loop to execute
 
     auto logTimeGuard = [&](const std::string command, int t) {
       if (t) {
@@ -2262,7 +2281,7 @@ int main(int argc, char* argv[])
 
     // loop for testing, single iteration in normal conditions
     for (int i = 0; i < nloop; i++) {
-      occRunNumber++;
+      incrementRunNumber();
       err = theReadout->start();
       if (err) {
         return err;

src/receiverFMQ.cxx

@@ -363,13 +363,15 @@ int main(int argc, const char** argv)
                   "STF:\n \
 		version: %d\n \
 		timeframeId: %d\n \
+		runNumber: %d\n \
 		systemId: %d\n \
 		feeId: %d\n \
 		equipmentId: %d\n \
 		linkId: %d\n\
 		lastTFMessage: %d\n",
                   (int)stf->version,
                   (int)stf->timeframeId,
+                  (int)stf->runNumber,
                   (int)stf->systemId,
                   (int)stf->feeId,
                   (int)stf->equipmentId,