Fix ScaledownIdletime = -1 to SuspendTime conversion

Fix ScaledownIdletime to SuspendTime conversion when ScaledownIdletime is set to -1 into cluster configuration. SuspendTime cannot be lower than -1, as it is not an accepted value by Slurm. Signed-off-by: Luca Carrogu <carrogu@amazon.com>
lukeseawalker · Dec 4, 2023 · 20a2a8d · 20a2a8d
1 parent c45b1d1
commit 20a2a8d
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 3 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -32,6 +32,7 @@ This file is used to list changes made in each version of the AWS ParallelCluste
 **BUG FIXES**
 - Fix inconsistent scaling configuration after cluster update rollback when modifying the list of instance types declared in the Compute Resources.
 - Fix users SSH keys generation when switching users without root privilege in clusters integrated with an external LDAP server through cluster configuration files.
+- Fix disabling Slurm power save mode when setting ScaledownIdletime = -1 
 
 3.7.2
 ------

diff --git a/...lelcluster-slurm/files/default/head_node_slurm/slurm/templates/slurm_parallelcluster.conf b/...lelcluster-slurm/files/default/head_node_slurm/slurm/templates/slurm_parallelcluster.conf
@@ -5,7 +5,7 @@
 {% set ns = namespace(has_static=false) %}
 
 SlurmctldHost={{ head_node_config.head_node_hostname }}({{ head_node_config.head_node_ip }})
-SuspendTime={{ scaling_config.ScaledownIdletime * 60 }}
+SuspendTime={{ -1 if scaling_config.ScaledownIdletime == -1 else scaling_config.ScaledownIdletime * 60 }}
 ResumeTimeout={{ compute_node_bootstrap_timeout }}
 {% if scaling_config.EnableMemoryBasedScheduling %}
 SelectTypeParameters=CR_CPU_Memory

diff --git a/...nerating_slurm_config_flexible_instance_types/expected_outputs/slurm_parallelcluster.conf b/...nerating_slurm_config_flexible_instance_types/expected_outputs/slurm_parallelcluster.conf
@@ -4,7 +4,7 @@
 # options
 
 SlurmctldHost=ip-1-0-0-0(ip.1.0.0.0)
-SuspendTime=600
+SuspendTime=-1
 ResumeTimeout=1600
 SelectTypeParameters=CR_CPU
 

diff --git a/...m_config_generator/test_generating_slurm_config_flexible_instance_types/sample_input.yaml b/...m_config_generator/test_generating_slurm_config_flexible_instance_types/sample_input.yaml
@@ -384,5 +384,5 @@ Scheduling:
       UseEc2Hostnames: false
     EnableMemoryBasedScheduling: false
     QueueUpdateStrategy: COMPUTE_FLEET_STOP
-    ScaledownIdletime: 10
+    ScaledownIdletime: -1
     Database: null