Refactored crash_gen tool, plus ensured proper termination upon trasntion failure

tests/pytest/crash_gen/db.py

@@ -78,7 +78,7 @@ class DbConn:
         if nRows != 1:
             raise taos.error.ProgrammingError(
                 "Unexpected result for query: {}, rows = {}".format(sql, nRows), 
-                (0x991 if nRows==0 else 0x992)
+                (CrashGenError.INVALID_EMPTY_RESULT if nRows==0 else CrashGenError.INVALID_MULTIPLE_RESULT)
             )
         if self.getResultRows() != 1 or self.getResultCols() != 1:
             raise RuntimeError("Unexpected result set for query: {}".format(sql))
@@ -349,7 +349,8 @@ class DbConnNative(DbConn):
 
     def execute(self, sql):
         if (not self.isOpen):
-            raise RuntimeError("Cannot execute database commands until connection is open")
+            raise CrashGenError(
+                "Cannot exec SQL unless db connection is open", CrashGenError.DB_CONNECTION_NOT_OPEN)
         Logging.debug("[SQL] Executing SQL: {}".format(sql))
         self._lastSql = sql
         nRows = self._tdSql.execute(sql)
@@ -360,8 +361,8 @@ class DbConnNative(DbConn):
 
     def query(self, sql):  # return rows affected
         if (not self.isOpen):
-            raise RuntimeError(
-                "Cannot query database until connection is open")
+            raise CrashGenError(
+                "Cannot query database until connection is open, restarting?", CrashGenError.DB_CONNECTION_NOT_OPEN)
         Logging.debug("[SQL] Executing SQL: {}".format(sql))
         self._lastSql = sql
         nRows = self._tdSql.query(sql)

tests/pytest/crash_gen/misc.py

@@ -3,14 +3,20 @@ import random
 import logging
 import os
 
+import taos
 
-class CrashGenError(Exception):
-    def __init__(self, msg=None, errno=None):
-        self.msg = msg
-        self.errno = errno
 
-    def __str__(self):
-        return self.msg
+class CrashGenError(taos.error.ProgrammingError):
+    INVALID_EMPTY_RESULT    = 0x991
+    INVALID_MULTIPLE_RESULT = 0x992
+    DB_CONNECTION_NOT_OPEN  = 0x993
+    # def __init__(self, msg=None, errno=None):
+    #     self.msg = msg
+    #     self.errno = errno
+
+    # def __str__(self):
+    #     return self.msg
+    pass
 
 
 class LoggingFilter(logging.Filter):
@@ -168,6 +174,7 @@ class Progress:
     SERVICE_RECONNECT_FAILURE   = 6
     SERVICE_START_NAP           = 7
     CREATE_TABLE_ATTEMPT        = 8
+    QUERY_GROUP_BY              = 9
 
     tokens = {
         STEP_BOUNDARY:      '.',
@@ -178,7 +185,8 @@ class Progress:
         SERVICE_RECONNECT_SUCCESS:  '.r>',
         SERVICE_RECONNECT_FAILURE:  '.xr>',
         SERVICE_START_NAP:           '_zz',
-        CREATE_TABLE_ATTEMPT: '_c',
+        CREATE_TABLE_ATTEMPT:       'c',
+        QUERY_GROUP_BY:             'g',
     }
 
     @classmethod

tests/pytest/crash_gen/service_manager.py

@@ -51,10 +51,12 @@ class TdeInstance():
     def prepareGcovEnv(cls, env):
         # Ref: https://gcc.gnu.org/onlinedocs/gcc/Cross-profiling.html
         bPath = cls._getBuildPath() # build PATH
-        numSegments = len(bPath.split('/')) - 1 # "/x/TDengine/build" should yield 3
-        numSegments = numSegments - 1 # DEBUG only
-        env['GCOV_PREFIX'] = bPath + '/svc_gcov'
+        numSegments = len(bPath.split('/')) # "/x/TDengine/build" should yield 3
+        # numSegments += 2 # cover "/src" after build
+        # numSegments = numSegments - 1 # DEBUG only
+        env['GCOV_PREFIX'] = bPath + '/src_s' # Server side source
         env['GCOV_PREFIX_STRIP'] = str(numSegments) # Strip every element, plus, ENV needs strings
+        # VERY VERY important note: GCOV data collection NOT effective upon SIG_KILL
         Logging.info("Preparing GCOV environement to strip {} elements and use path: {}".format(
             numSegments, env['GCOV_PREFIX'] ))
 
@@ -258,14 +260,15 @@ class TdeSubProcess:
         TdeInstance.prepareGcovEnv(myEnv)
 
         # print(myEnv)
-        # print(myEnv.items())
+        # print("Starting TDengine with env: ", myEnv.items())
         # print("Starting TDengine via Shell: {}".format(cmdLineStr))
 
         useShell = True    
         self.subProcess = subprocess.Popen(
-            ' '.join(cmdLine) if useShell else cmdLine,
-            shell=useShell,
-            # svcCmdSingle, shell=True, # capture core dump?
+            # ' '.join(cmdLine) if useShell else cmdLine,
+            # shell=useShell,
+            ' '.join(cmdLine),
+            shell=True,
             stdout=subprocess.PIPE,
             stderr=subprocess.PIPE,
             # bufsize=1, # not supported in binary mode
@@ -273,7 +276,8 @@ class TdeSubProcess:
             env=myEnv
             )  # had text=True, which interferred with reading EOF
 
-    STOP_SIGNAL = signal.SIGKILL # What signal to use (in kill) to stop a taosd process?
+    STOP_SIGNAL = signal.SIGKILL # signal.SIGKILL/SIGINT # What signal to use (in kill) to stop a taosd process?
+    SIG_KILL_RETCODE = 137 # ref: https://stackoverflow.com/questions/43268156/process-finished-with-exit-code-137-in-pycharm
 
     def stop(self):
         """
@@ -320,8 +324,12 @@ class TdeSubProcess:
         retCode = self.subProcess.returncode # should always be there
         # May throw subprocess.TimeoutExpired exception above, therefore
         # The process is guranteed to have ended by now
-        self.subProcess = None        
-        if retCode != 0: # != (- signal.SIGINT):
+        self.subProcess = None       
+        if retCode == self.SIG_KILL_RETCODE:
+            Logging.info("TSP.stop(): sub proc KILLED, as expected")
+        elif retCode == (- self.STOP_SIGNAL):
+            Logging.info("TSP.stop(), sub process STOPPED, as expected")
+        elif retCode != 0: # != (- signal.SIGINT):
             Logging.error("TSP.stop(): Failed to stop sub proc properly w/ SIG {}, retCode={}".format(
                 self.STOP_SIGNAL, retCode))
         else: