fix: handling of remaining jobs when using --keep-going (#1693)

* fix: proper handling of remaining jobs in case of keepgoing behavior

* syntax

snakemake/scheduler.py

@@ -404,7 +404,9 @@ def remaining_jobs(self):
         return [
             job
             for job in self.dag.needrun_jobs()
-            if job not in self.running and not self.dag.finished(job)
+            if job not in self.running
+            and not self.dag.finished(job)
+            and job not in self.failed
         ]
 
     def schedule(self):
@@ -451,7 +453,7 @@ def schedule(self):
                         logger.error(_ERROR_MSG_FINAL)
                     # we still have unfinished jobs. this is not good. direct
                     # user to github issue
-                    if self.remaining_jobs:
+                    if self.remaining_jobs and not self.keepgoing:
                         logger.error(_ERROR_MSG_ISSUE_823)
                         logger.error(
                             "Remaining jobs:\n"
@@ -609,15 +611,13 @@ def _handle_error(self, job):
         # attempt starts counting from 1, but the first attempt is not
         # a restart, hence we subtract 1.
         if job.restart_times > job.attempt - 1:
-            logger.info("Trying to restart job {}.".format(self.dag.jobid(job)))
+            logger.info(f"Trying to restart job {self.dag.jobid(job)}.")
             job.attempt += 1
             # add job to those being ready again
             self.dag._ready_jobs.add(job)
         else:
             self._errors = True
             self.failed.add(job)
-            if self.keepgoing:
-                logger.info("Job failed, going on with independent jobs.")
 
     def exit_gracefully(self, *args):
         with self._lock: