postmaster: Move code to switch into FatalError state into function

anarazel · anarazel · commit 8edd8c77c88e · 2025-01-24T17:00:10.000-05:00
There are two places switching to FatalError mode, behaving somewhat differently. An upcoming commit will introduce a third. That doesn't seem seem like a good idea. This commit just moves the FatalError related code from HandleChildCrash() into its own function, a subsequent commit will evolve the state machine change to be suitable for other callers. Reviewed-by: Bertrand Drouvot <bertranddrouvot.pg@gmail.com> Discussion: https://postgr.es/m/kgng5nrvnlv335evmsuvpnh354rw7qyazl73kdysev2cr2v5zu@m3cfzxicm5kp
diff --git a/src/backend/postmaster/postmaster.c b/src/backend/postmaster/postmaster.c
@@ -2674,40 +2674,35 @@ CleanupBackend(PMChild *bp,
 }
 
 /*
- * HandleChildCrash -- cleanup after failed backend, bgwriter, checkpointer,
- * walwriter, autovacuum, archiver, slot sync worker, or background worker.
+ * Transition into FatalError state, in response to something bad having
+ * happened. Commonly the caller will have logged the reason for entering
+ * FatalError state.
  *
- * The objectives here are to clean up our local state about the child
- * process, and to signal all other remaining children to quickdie.
- *
- * The caller has already released its PMChild slot.
+ * This should only be called when not already in FatalError or
+ * ImmediateShutdown state.
  */
 static void
-HandleChildCrash(int pid, int exitstatus, const char *procname)
+HandleFatalError(QuitSignalReason reason, bool consider_sigabrt)
 {
-	/*
-	 * We only log messages and send signals if this is the first process
-	 * crash and we're not doing an immediate shutdown; otherwise, we're only
-	 * here to update postmaster's idea of live processes.  If we have already
-	 * signaled children, nonzero exit status is to be expected, so don't
-	 * clutter log.
-	 */
-	if (FatalError || Shutdown == ImmediateShutdown)
-		return;
+	int			sigtosend;
 
-	LogChildExit(LOG, procname, pid, exitstatus);
-	ereport(LOG,
-			(errmsg("terminating any other active server processes")));
-	SetQuitSignalReason(PMQUIT_FOR_CRASH);
+	Assert(!FatalError);
+	Assert(Shutdown != ImmediateShutdown);
+
+	SetQuitSignalReason(reason);
+
+	if (consider_sigabrt && send_abort_for_crash)
+		sigtosend = SIGABRT;
+	else
+		sigtosend = SIGQUIT;
 
 	/*
-	 * Signal all other child processes to exit.  The crashed process has
-	 * already been removed from ActiveChildList.
+	 * Signal all other child processes to exit.
 	 *
 	 * We could exclude dead-end children here, but at least when sending
 	 * SIGABRT it seems better to include them.
 	 */
-	TerminateChildren(send_abort_for_crash ? SIGABRT : SIGQUIT);
+	TerminateChildren(sigtosend);
 
 	FatalError = true;
 
@@ -2727,6 +2722,39 @@ HandleChildCrash(int pid, int exitstatus, const char *procname)
 		AbortStartTime = time(NULL);
 }
 
+/*
+ * HandleChildCrash -- cleanup after failed backend, bgwriter, checkpointer,
+ * walwriter, autovacuum, archiver, slot sync worker, or background worker.
+ *
+ * The objectives here are to clean up our local state about the child
+ * process, and to signal all other remaining children to quickdie.
+ *
+ * The caller has already released its PMChild slot.
+ */
+static void
+HandleChildCrash(int pid, int exitstatus, const char *procname)
+{
+	/*
+	 * We only log messages and send signals if this is the first process
+	 * crash and we're not doing an immediate shutdown; otherwise, we're only
+	 * here to update postmaster's idea of live processes.  If we have already
+	 * signaled children, nonzero exit status is to be expected, so don't
+	 * clutter log.
+	 */
+	if (FatalError || Shutdown == ImmediateShutdown)
+		return;
+
+	LogChildExit(LOG, procname, pid, exitstatus);
+	ereport(LOG,
+			(errmsg("terminating any other active server processes")));
+
+	/*
+	 * Switch into error state. The crashed process has already been removed
+	 * from ActiveChildList.
+	 */
+	HandleFatalError(PMQUIT_FOR_CRASH, true);
+}
+
 /*
  * Log the death of a child process.
  */
