leave a comment about 0 capacity unused partitions

Azure · Oct 21, 2024 · 4adee8b · 4adee8b
1 parent 405ef5f
commit 4adee8b
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 9 deletions.
diff --git a/slurm/src/slurmcc/cli.py b/slurm/src/slurmcc/cli.py
@@ -803,18 +803,18 @@ def _partitions(
             threads = 1
         def_mem_per_cpu = memory // cpus
 
-
+        comment_out = ""
+        if max_count <= 0:
+            writer.write(f"# The following partition has no capacity! {partition.name} - {partition.nodearray} - {partition.machine_type} \n")
+            comment_out = "# "
+
         writer.write(
-            "PartitionName={} Nodes={} Default={} DefMemPerCPU={} MaxTime=INFINITE State=UP\n".format(
-                partition.name, partition.node_list, default_yn, def_mem_per_cpu
-            )
+            f"{comment_out}PartitionName={partition.name} Nodes={partition.node_list} Default={default_yn} DefMemPerCPU={def_mem_per_cpu} MaxTime=INFINITE State=UP\n"
         )
 
         state = "CLOUD" if autoscale else "FUTURE"
         writer.write(
-            "Nodename={} Feature=cloud STATE={} CPUs={} ThreadsPerCore={} RealMemory={}".format(
-                node_list, state, cpus, threads, memory
-            )
+           f"{comment_out}Nodename={node_list} Feature=cloud STATE={state} CPUs={cpus} ThreadsPerCore={threads} RealMemory={memory}"
         )
 
         if partition.gpu_count:

diff --git a/slurm/src/slurmcc/partition.py b/slurm/src/slurmcc/partition.py
@@ -104,6 +104,9 @@ def _slurm_nodes(cls) -> List[Dict[str, str]]:
     @property
     def node_list(self) -> str:
         if not self.dynamic_config:
+            static_nodes = self._static_all_nodes()
+            if not static_nodes:
+                return ""
             return slutil.to_hostlist(self._static_all_nodes())
         # with dynamic nodes, we only look at those defined in the partition
         if not self.__dynamic_node_list_cache:
@@ -349,10 +352,10 @@ def fetch_partitions(
 
             if limits.max_count <= 0:
                 logging.info(
-                    "Bucket has a max_count <= 0, defined for machinetype=='%s'. Skipping",
+                    "Bucket has a max_count <= 0, defined for machinetype=='%s'.",
                     machine_type,
                 )
-                continue
+                # keep this partition around, but we will ignore it when generating later.
 
             max_scaleset_size = buckets[0].max_placement_group_size