Several updates and fixes (version, process, doc)

fix typo on SLURM_VERSION SLURM_VERSION update slurm to version 24.05 add documentation on how to create openssl private key for the ssh server add missing cgroup.conf and disable cgroup usage from: @akram moved patch over from old kincl/openshift-hpc-batch-demo repo
naps-product-sa · Aug 30, 2024 · c0b9580 · c0b9580
1 parent 39f5408
commit c0b9580
Show file tree

Hide file tree

Showing 8 changed files with 60 additions and 16 deletions.
diff --git a/slurm/README.md b/slurm/README.md
@@ -14,7 +14,7 @@ oc apply -k manifests/
 oc -n slurm-system create secret generic munge-key --from-file=munge.key=<(dd status=none if=/dev/urandom bs=1 count=128)
 ```
 
-### Build images
+### Create image builds and build images
 
 ```
 oc new-build --name munge --binary
@@ -54,13 +54,39 @@ sleep 10
 
 ## containerssh
 
-TODO: replace authconfig python with OPA example: https://github.com/ContainerSSH/examples/tree/main/opa/containerssh
+```
+oc new-project containerssh
+```
+
+Set your user ssh public key into the authentication app (hardcoded)
+
+```
+sed -i -e "s#ssh-rsa.*#$(cat ~/.ssh/id_rsa.pub)#" manifests/containerssh/authconfig/app.py
+```
 
+TODO: Automate or script this part
+Generate host key and create a secret for it
+
+```
+openssl rsa -in mykey.pem -pubout > mykey.pub
+oc create secret generic -n containerssh  containerssh-hostkey --from-file=host.key=mykey.pem
+```
 ```
 $ oc apply -k manifests/containerssh
 $ openssl genrsa | kubectl create secret generic -n containerssh containerssh-hostkey --from-file=host.key=/dev/stdin
 ```
 
+## Checks
+
+Check thats pods in projects `slurm-system` and `containerssh` are up and running:
+
+```
+oc get pods -n containerssh
+oc get pods -n slurm-system
+```
+
+You are ready for the demo and you can return to [the main demo README](../README.md) 
+
 ### Login
 
 ```
@@ -70,4 +96,4 @@ containerssh   LoadBalancer   172.30.147.21   a9cf4fabd2d9b49659d8af106ea30536-1
 
 $ ssh -o HostKeyAlgorithms=+ssh-rsa -o PubkeyAcceptedAlgorithms=+ssh-rsa -p 2222 jason@a9cf4fabd2d9b49659d8af106ea30536-1220395463.us-east-2.elb.amazonaws.com
 bash-5.1$
-```
+```
diff --git a/slurm/images/slurm/Dockerfile b/slurm/images/slurm/Dockerfile
@@ -1,5 +1,5 @@
 FROM ghcr.io/naps-product-sa/openshift-batch/munge:latest as build
-ARG SLURM_VERSION=23.02.4
+ARG SLURM_VERSION=24.05.3
 
 RUN dnf install --enablerepo=crb -y xz gcc procps-ng bzip2 perl git autoconf automake libtool diffutils jansson \
     openssl-devel bzip2-devel zlib-devel jansson-devel http-parser-devel json-c-devel perl-devel

diff --git a/slurm/manifests/configs/cgroup.conf b/slurm/manifests/configs/cgroup.conf
@@ -0,0 +1 @@
+CgroupPlugin=disabled
diff --git a/slurm/manifests/configs/slurm.conf b/slurm/manifests/configs/slurm.conf
@@ -129,7 +129,7 @@ SlurmctldDebug=debug
 #SlurmctldLogFile=
 SlurmctldParameters=cloud_reg_addrs
 SlurmdDebug=debug
-#SlurmdLogFile=
+SlurmdLogFile=/tmp/slurmd.log
 #SlurmSchedLogFile=
 #SlurmSchedLogLevel=
 #DebugFlags=

diff --git a/slurm/manifests/containerssh/configs/config.yaml b/slurm/manifests/containerssh/configs/config.yaml
@@ -26,7 +26,7 @@ kubernetes:
       serviceAccountName: slurm
       initContainers:
       - name: fix-munge-paths
-        image: ghcr.io/naps-product-sa/openshift-batch/munge:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/munge:latest
         command:
         - sh
         - -c
@@ -40,7 +40,7 @@ kubernetes:
         - name: munge-socket
           mountPath: /run/munge
       - name: add-jobs
-        image: ghcr.io/naps-product-sa/openshift-batch/login:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/login:latest
         command:
         - sh
         - -c
@@ -69,7 +69,7 @@ kubernetes:
           mountPath: /home
       containers:
       - name: shell
-        image: ghcr.io/naps-product-sa/openshift-batch/login:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/login:latest
         securityContext:
           runAsUser: 1000
           runAsGroup: 1000
@@ -82,7 +82,7 @@ kubernetes:
         - name: shared
           mountPath: /home
       - name: munge
-        image: ghcr.io/naps-product-sa/openshift-batch/munge:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/munge:latest
         command:
         - /sbin/munged
         - -F

diff --git a/slurm/manifests/kustomization.yaml b/slurm/manifests/kustomization.yaml
@@ -23,3 +23,8 @@ configMapGenerator:
     - configs/slurm.conf
   options:
     disableNameSuffixHash: true
+    labels:
+      app: slurm
+- name: cgroup-conf
+  files:
+    - configs/cgroup.conf
diff --git a/slurm/manifests/statefulset-compute.yaml b/slurm/manifests/statefulset-compute.yaml
@@ -60,7 +60,7 @@ spec:
             topologyKey: "kubernetes.io/hostname"
       initContainers:
       - name: fix-munge-paths
-        image: ghcr.io/naps-product-sa/openshift-batch/munge:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/munge:latest
         command:
         - sh
         - -c
@@ -75,7 +75,7 @@ spec:
           mountPath: /run/munge
       containers:
       - name: slurm
-        image: ghcr.io/naps-product-sa/openshift-batch/slurm:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/slurm:latest
         command:
         - /sbin/slurmd
         - -D
@@ -104,10 +104,13 @@ spec:
         - name: slurm-conf
           mountPath: /etc/slurm.conf
           subPath: slurm.conf
+        - name: cgroup-conf
+          mountPath: /etc/cgroup.conf
+          subPath: cgroup.conf
         - name: shared
           mountPath: /home
       - name: munge
-        image: ghcr.io/naps-product-sa/openshift-batch/munge:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/munge:latest
         command:
         - /bin/bash
         - /scripts/munged-container.sh
@@ -131,6 +134,9 @@ spec:
       - name: slurm-conf
         configMap:
           name: slurm-conf
+      - name: cgroup-conf
+        configMap:
+          name: cgroup-conf
 
       - name: shared
         persistentVolumeClaim:

diff --git a/slurm/manifests/statefulset-head.yaml b/slurm/manifests/statefulset-head.yaml
@@ -34,7 +34,7 @@ spec:
         fsGroupChangePolicy: Always
       initContainers:
       - name: fix-munge-paths
-        image: ghcr.io/naps-product-sa/openshift-batch/munge:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/munge:latest
         command:
         - sh
         - -c
@@ -49,7 +49,7 @@ spec:
           mountPath: /run/munge
       containers:
       - name: slurm
-        image: ghcr.io/naps-product-sa/openshift-batch/slurm:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/slurm:latest
         command:
         - /sbin/slurmctld
         - -D
@@ -67,7 +67,7 @@ spec:
         - name: shared
           mountPath: /home
       - name: exporter
-        image: ghcr.io/naps-product-sa/openshift-batch/slurm:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/slurm:latest
         command:
         - /usr/bin/prometheus-slurm-exporter
         securityContext:
@@ -79,8 +79,11 @@ spec:
         - name: slurm-conf
           mountPath: /etc/slurm.conf
           subPath: slurm.conf
+        - name: cgroup-conf
+          mountPath: /etc/cgroup.conf
+          subPath: cgroup.conf
       - name: munge
-        image: ghcr.io/naps-product-sa/openshift-batch/munge:latest
+        image: image-registry.openshift-image-registry.svc:5000/slurm-system/munge:latest
         command:
         - /sbin/munged
         - -F
@@ -98,6 +101,9 @@ spec:
       - name: slurm-conf
         configMap:
           name: slurm-conf
+      - name: cgroup-conf
+        configMap:
+          name: cgroup-conf
 
       - name: munge-key
         emptyDir: {}