fix: improved error handling for cluster status scripts and smarter job selector choice in case of cluster submission (use greedy for single jobs).

snakemake/executors/__init__.py

@@ -712,7 +712,7 @@ def __init__(
         self.active_jobs = list()
         self.lock = threading.Lock()
         self.wait = True
-        self.wait_thread = threading.Thread(target=self._wait_for_jobs)
+        self.wait_thread = threading.Thread(target=self._wait_thread)
         self.wait_thread.daemon = True
         self.wait_thread.start()
 
@@ -722,6 +722,12 @@ def __init__(
             max_calls=self.max_status_checks_per_second, period=1
         )
 
+    def _wait_thread(self):
+        try:
+            self._wait_for_jobs()
+        except Exception as e:
+            self.workflow.scheduler.executor_error_callback(e)
+
     def shutdown(self):
         with self.lock:
             self.wait = False
@@ -1065,21 +1071,18 @@ def _wait_for_jobs(self):
         success = "success"
         failed = "failed"
         running = "running"
+        status_cmd_kills = set()
         if self.statuscmd is not None:
 
-            def job_status(job):
+            def job_status(job, valid_returns=["running", "success", "failed"]):
                 try:
                     # this command shall return "success", "failed" or "running"
-                    return (
-                        subprocess.check_output(
-                            "{statuscmd} {jobid}".format(
-                                jobid=job.jobid, statuscmd=self.statuscmd
-                            ),
-                            shell=True,
-                        )
-                        .decode()
-                        .split("\n")[0]
-                    )
+                    ret = subprocess.check_output(
+                        "{statuscmd} {jobid}".format(
+                            jobid=job.jobid, statuscmd=self.statuscmd
+                        ),
+                        shell=True,
+                    ).decode()
                 except subprocess.CalledProcessError as e:
                     if e.returncode < 0:
                         # Ignore SIGINT and all other issues due to signals
@@ -1088,13 +1091,31 @@ def job_status(job):
                         # snakemake.
                         # Snakemake will handle the signal in
                         # the main process.
-                        pass
+                        status_cmd_kills.add(e.returncode)
+                        if len(status_cmd_kills) > 10:
+                            logger.info(
+                                "Cluster status command {} was killed >10 times with signal(s) {} "
+                                "(if this happens unexpectedly during your workflow execution, "
+                                "have a closer look.).".format(
+                                    self.statuscmd, ",".join(status_cmd_kills)
+                                )
+                            )
+                            status_cmd_kills.clear()
                     else:
                         raise WorkflowError(
                             "Failed to obtain job status. "
                             "See above for error message."
                         )
 
+                ret = ret.strip().split("\n")
+                if len(ret) != 1 or ret[0] not in valid_returns:
+                    raise WorkflowError(
+                        "Cluster status command {} returned {} but just a single line with one of {} is expected.".format(
+                            self.statuscmd, "\\n".join(ret), ",".join(valid_returns)
+                        )
+                    )
+                return ret[0]
+
         else:
 
             def job_status(job):

snakemake/scheduler.py

@@ -140,8 +140,10 @@ def __init__(
         self._lock = threading.Lock()
 
         self._errors = False
+        self._executor_error = None
         self._finished = False
         self._job_queue = None
+        self._last_job_selection_empty = False
         self._submit_callback = self._noop
         self._finish_callback = partial(
             self._proceed,
@@ -391,6 +393,12 @@ def __init__(
             pass
         self._open_jobs.release()
 
+    def executor_error_callback(self, exception):
+        with self._lock:
+            self._executor_error = exception
+            # next scheduling round to catch and raise error
+            self._open_jobs.release()
+
     @property
     def stats(self):
         try:
@@ -435,16 +443,20 @@ def schedule(self):
                     needrun = set(self.open_jobs)
                     running = list(self.running)
                     errors = self._errors
+                    executor_error = self._executor_error
                     user_kill = self._user_kill
 
                 # handle errors
-                if user_kill or (not self.keepgoing and errors):
+                if user_kill or (not self.keepgoing and errors) or executor_error:
                     if user_kill == "graceful":
                         logger.info(
                             "Will exit after finishing " "currently running jobs."
                         )
 
-                    if not running:
+                    if executor_error:
+                        print_exception(executor_error, self.workflow.linemaps)
+
+                    if executor_error or not running:
                         logger.info("Shutting down, this might take some time.")
                         self._executor.shutdown()
                         if not user_kill:
@@ -474,7 +486,11 @@ def schedule(self):
                         "Ready jobs ({}):\n\t".format(len(needrun))
                         + "\n\t".join(map(str, needrun))
                     )
+
+                    if not self._last_job_selection_empty:
+                        logger.info("Select jobs to execute...")
                     run = self.job_selector(needrun)
+                    self._last_job_selection_empty = not run
 
                     logger.debug(
                         "Selected jobs ({}):\n\t".format(len(run))
@@ -624,7 +640,11 @@ def job_selector_ilp(self, jobs):
         from pulp import lpSum
         from stopit import ThreadingTimeout as Timeout, TimeoutException
 
-        logger.info("Select jobs to execute...")
+        if len(jobs) == 1:
+            logger.debug(
+                "Using greedy selector because only single job has to be scheduled."
+            )
+            return self.job_selector_greedy(jobs)
 
         with self._lock:
             if not self.resources["_cores"]:
@@ -895,8 +915,14 @@ def job_reward(self, job):
             temp_size = 0
             input_size = 0
         else:
-            temp_size = self.dag.temp_size(job)
-            input_size = job.inputsize
+            try:
+                temp_size = self.dag.temp_size(job)
+                input_size = job.inputsize
+            except FileNotFoundError:
+                # If the file is not yet present, this shall not affect the
+                # job selection.
+                temp_size = 0
+                input_size = 0
 
         # Usually, this should guide the scheduler to first schedule all jobs
         # that remove the largest temp file, then the second largest and so on.

snakemake/workflow.py

@@ -217,6 +217,7 @@ def __init__(
         self.check_envvars = check_envvars
         self.max_threads = max_threads
         self.all_temp = all_temp
+        self.scheduler = None
 
         _globals = globals()
         _globals["workflow"] = self
@@ -953,7 +954,7 @@ def files(items):
             self.persistence.conda_cleanup_envs()
             return True
 
-        scheduler = JobScheduler(
+        self.scheduler = JobScheduler(
             self,
             dag,
             local_cores=local_cores,
@@ -1053,7 +1054,7 @@ def files(items):
         if not dryrun and not no_hooks:
             self._onstart(logger.get_logfile())
 
-        success = scheduler.schedule()
+        success = self.scheduler.schedule()
 
         if not immediate_submit and not dryrun:
             dag.cleanup_workdir()
@@ -1069,7 +1070,7 @@ def files(items):
                 logger.remove_logfile()
             else:
                 if stats:
-                    scheduler.stats.to_json(stats)
+                    self.scheduler.stats.to_json(stats)
                 logger.logfile_hint()
             if not dryrun and not no_hooks:
                 self._onsuccess(logger.get_logfile())

tests/test_cluster_statusscript/Snakefile.nonstandard

@@ -0,0 +1,32 @@
+from snakemake import shell
+
+chromosomes = [1,2,3,4,5]
+
+envvars:
+    "TESTVAR"
+
+
+
+rule all:
+	input: 'test.predictions', 'test.2.inter2'
+
+rule compute1:
+	input: '{name}.in'
+	output: ['{name}.%s.inter'%c for c in chromosomes]
+        params: prefix="{name}"
+	run:
+		for out in output:
+			shell('(cat {input[0]} && echo "Part {out}") > {out}')
+
+rule compute2:
+	input: '{name}.{chromosome}.inter'
+	output: '{name}.{chromosome}.inter2'
+	params: test="a=b"
+        threads: workflow.cores * 0.5
+	shell: 'echo copy; cp {input[0]} {output[0]}'
+
+rule gather:
+	input: ['{name}.%s.inter2'%c for c in chromosomes]
+	output: '{name}.predictions'
+	run:
+		shell('cat {} > {}'.format(' '.join(input), output[0]))

tests/test_cluster_statusscript/expected-results/test.1.inter

@@ -0,0 +1,2 @@
+testz0r
+Part test.1.inter

tests/test_cluster_statusscript/expected-results/test.1.inter2

@@ -0,0 +1,2 @@
+testz0r
+Part test.1.inter

tests/test_cluster_statusscript/expected-results/test.2.inter

@@ -0,0 +1,2 @@
+testz0r
+Part test.2.inter

tests/test_cluster_statusscript/expected-results/test.2.inter2

@@ -0,0 +1,2 @@
+testz0r
+Part test.2.inter

tests/test_cluster_statusscript/expected-results/test.3.inter

@@ -0,0 +1,2 @@
+testz0r
+Part test.3.inter

tests/test_cluster_statusscript/expected-results/test.3.inter2

@@ -0,0 +1,2 @@
+testz0r
+Part test.3.inter

tests/test_cluster_statusscript/expected-results/test.predictions

@@ -0,0 +1,10 @@
+testz0r
+Part test.1.inter
+testz0r
+Part test.2.inter
+testz0r
+Part test.3.inter
+testz0r
+Part test.4.inter
+testz0r
+Part test.5.inter

tests/test_cluster_statusscript/qsub

@@ -0,0 +1,7 @@
+#!/bin/bash
+echo `date` >> qsub.log
+tail -n1 $1 >> qsub.log
+# simulate printing of job id by a random number
+echo $RANDOM
+cat $1 >> qsub.log
+sh $1

tests/test_cluster_statusscript/status.sh

@@ -0,0 +1 @@
+echo success

tests/test_cluster_statusscript/test.in

@@ -0,0 +1 @@
+testz0r

tests/tests.py

@@ -104,6 +104,17 @@ def test14():
     run(dpath("test14"), snakefile="Snakefile.nonstandard", cluster="./qsub")
 
 
+@skip_on_windows
+def test_cluster_statusscript():
+    os.environ["TESTVAR"] = "test"
+    run(
+        dpath("test_cluster_statusscript"),
+        snakefile="Snakefile.nonstandard",
+        cluster="./qsub",
+        cluster_status="./status.sh",
+    )
+
+
 def test15():
     run(dpath("test15"))