diff --git a/src/lpcjobqueue/cluster.py b/src/lpcjobqueue/cluster.py
index 8318332..7072e92 100644
--- a/src/lpcjobqueue/cluster.py
+++ b/src/lpcjobqueue/cluster.py
@@ -54,6 +54,7 @@ def __init__(
         **base_class_kwargs,
     ):
         image = self.container_prefix + image
+        image = os.path.realpath(image)
         if ship_env:
             base_class_kwargs["python"] = f"{self.env_name}/bin/python"
             base_class_kwargs.setdefault(
@@ -93,20 +94,10 @@ async def start(self):
         job = htcondor.Submit(job)
 
         def sub():
-            try:
-                classads = []
-                with SCHEDD().transaction() as txn:
-                    cluster_id = job.queue(txn, ad_results=classads)
-
-                logger.debug(f"ClassAds for job {cluster_id}: {classads}")
-                SCHEDD().spool(classads)
-                return cluster_id
-            except htcondor.HTCondorInternalError as ex:
-                logger.error(str(ex))
-                return None
-            except htcondor.HTCondorIOError as ex:
-                logger.error(str(ex))
-                return None
+            result = SCHEDD().submit(job, spool=True)
+            cluster_id = result.cluster()
+            SCHEDD().spool(list(job.jobs(clusterid=cluster_id)))
+            return cluster_id
 
         self.job_id = await asyncio.get_event_loop().run_in_executor(SCHEDD_POOL, sub)
         if self.job_id:
@@ -147,7 +138,7 @@ def check_gone():
                     SCHEDD_POOL, check_gone
                 )
             except RuntimeError as ex:
-                if str(ex) == "cannot schedule new futures after interpreter shutdown":
+                if str(ex) == "cannot schedule new futures after shutdown":
                     logger.info(f"Thread pool lost while checking worker {self.name} job {self.job_id}")
                     # We're not going to be able to do anything async now
                     self.status = Status.undefined