make --syslog errors non fatal

Podman will always pass down --syslog to conmon since 13c2aca21.
However there systems without syslog running, likely in container
setups. As reported in this was already a problem before when debug
level is used. Then conmon will pass down --syslog back to the podman
container cleanup command causing it to fail without doing anything.
Given that I think it is better to just ignore the error and log it on
debug level, we need to make sure cleanup works consistently.

[NO NEW TESTS NEEDED]

Fixes #19075

Signed-off-by: Paul Holzinger <pholzing@redhat.com>

cmd/podman/syslog_common.go

@@ -4,9 +4,7 @@
 package main
 
 import (
-	"fmt"
 	"log/syslog"
-	"os"
 
 	"github.com/sirupsen/logrus"
 	logrusSyslog "github.com/sirupsen/logrus/hooks/syslog"
@@ -19,10 +17,8 @@ func syslogHook() {
 
 	hook, err := logrusSyslog.NewSyslogHook("", "", syslog.LOG_INFO, "")
 	if err != nil {
-		fmt.Fprint(os.Stderr, "Failed to initialize syslog hook: "+err.Error())
-		os.Exit(1)
-	}
-	if err == nil {
+		logrus.Debug("Failed to initialize syslog hook: " + err.Error())
+	} else {
 		logrus.AddHook(hook)
 	}
 }