Slurm version update and rest service add

This commit is contained in:
Aditya Ujeniya 2024-10-22 20:27:09 +02:00
parent 2d15d513c6
commit 2b68597724
14 changed files with 99 additions and 101 deletions

View File

@ -69,8 +69,6 @@ services:
build:
context: ./slurm/controller
privileged: true
ports:
- "6817:6817"
volumes:
- ${DATADIR}/slurm/home:/home
- ${DATADIR}/slurm/secret:/.secret
@ -78,6 +76,8 @@ services:
- /etc/timezone:/etc/timezone:ro
- /etc/localtime:/etc/localtime:ro
- ${DATADIR}/slurm/state:/var/lib/slurm/d
ports:
- "6817:6817"
slurmdbd:
container_name: slurmdbd
@ -88,14 +88,15 @@ services:
- mariadb
- slurmctld
privileged: true
ports:
- "6819:6819"
volumes:
- ${DATADIR}/slurm/home:/home
- ${DATADIR}/slurm/secret:/.secret
- ./slurm/database/slurmdbd.conf:/home/config/slurmdbd.conf
- /etc/timezone:/etc/timezone:ro
- /etc/localtime:/etc/localtime:ro
- ${DATADIR}/slurm/state:/var/lib/slurm/d
ports:
- "6819:6819"
node01:
container_name: node01
@ -114,3 +115,21 @@ services:
- /etc/localtime:/etc/localtime:ro
ports:
- "6818:6818"
slurmrestd:
container_name: slurmrestd
hostname: slurmrestd
build:
context: ./slurm/rest
depends_on:
- slurmctld
privileged: true
volumes:
- ${DATADIR}/slurm/home:/home
- ${DATADIR}/slurm/secret:/.secret
- ./slurm/controller/slurm.conf:/home/config/slurm.conf
- ./slurm/rest/slurmrestd.conf:/home/config/slurmrestd.conf
- /etc/timezone:/etc/timezone:ro
- /etc/localtime:/etc/localtime:ro
ports:
- "6820:6820"

View File

@ -88,6 +88,7 @@ if [ ! -f docker-compose.yml ]; then
fi
docker-compose down
docker-compose down --remove-orphans
cd slurm/base/
make

View File

@ -1,9 +1,11 @@
FROM rockylinux:8
LABEL org.opencontainers.image.authors="jan.eitzinger@fau.de"
ENV SLURM_VERSION=22.05.6
ENV SLURM_VERSION=24.05.3
ENV HTTP_PARSER_VERSION=2.8.0
RUN yum install -y https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm
RUN ARCH=$(uname -m) && yum install -y https://rpmfind.net/linux/almalinux/8.10/PowerTools/x86_64/os/Packages/http-parser-devel-2.8.0-9.el8.$ARCH.rpm
RUN groupadd -g 981 munge \
&& useradd -m -c "MUNGE Uid 'N' Gid Emporium" -d /var/lib/munge -u 981 -g munge -s /sbin/nologin munge \
@ -17,20 +19,21 @@ RUN yum install -y munge munge-libs rng-tools \
openssh-server openssh-clients dbus-devel \
pam-devel numactl numactl-devel hwloc sudo \
lua readline-devel ncurses-devel man2html \
libibmad libibumad rpm-build perl-ExtUtils-MakeMaker.noarch rpm-build make wget
autoconf automake json-c-devel \
libibmad libibumad rpm-build perl-ExtUtils-MakeMaker.noarch rpm-build make wget
RUN dnf --enablerepo=powertools install -y munge-devel rrdtool-devel lua-devel hwloc-devel mariadb-server mariadb-devel
RUN mkdir -p /usr/local/slurm-tmp \
&& cd /usr/local/slurm-tmp \
&& wget https://download.schedmd.com/slurm/slurm-${SLURM_VERSION}.tar.bz2 \
&& rpmbuild -ta slurm-${SLURM_VERSION}.tar.bz2
&& rpmbuild -ta --with slurmrestd slurm-${SLURM_VERSION}.tar.bz2
RUN ARCH=$(uname -m) \
&& yum -y --nogpgcheck localinstall \
/root/rpmbuild/RPMS/$ARCH/slurm-${SLURM_VERSION}-1.el8.$ARCH.rpm \
/root/rpmbuild/RPMS/$ARCH/slurm-perlapi-${SLURM_VERSION}-1.el8.$ARCH.rpm \
/root/rpmbuild/RPMS/$ARCH/slurm-slurmctld-${SLURM_VERSION}-1.el8.$ARCH.rpm
/root/rpmbuild/RPMS/$ARCH/slurm-${SLURM_VERSION}*.$ARCH.rpm \
/root/rpmbuild/RPMS/$ARCH/slurm-perlapi-${SLURM_VERSION}*.$ARCH.rpm \
/root/rpmbuild/RPMS/$ARCH/slurm-slurmctld-${SLURM_VERSION}*.$ARCH.rpm
VOLUME ["/home", "/.secret"]
# 22: SSH
@ -38,4 +41,5 @@ VOLUME ["/home", "/.secret"]
# 6817: SlurmCtlD
# 6818: SlurmD
# 6819: SlurmDBD
EXPOSE 22 6817 6818 6819
# 6820: SlurmRestD
EXPOSE 22 6817 6818 6819 6820

View File

@ -1,6 +1,6 @@
include ../../.env
IMAGE = clustercockpit/slurm.base
SLURM_VERSION = 24.05.3
.PHONY: build clean
build:

View File

@ -1,4 +1,4 @@
FROM clustercockpit/slurm.base:22.05.6
FROM clustercockpit/slurm.base:24.05.3
LABEL org.opencontainers.image.authors="jan.eitzinger@fau.de"
# clean up

View File

@ -3,6 +3,7 @@ set -e
# Determine the system architecture dynamically
ARCH=$(uname -m)
SLURM_VERSION="24.05.3"
_delete_secrets() {
if [ -f /.secret/munge.key ]; then
@ -90,11 +91,12 @@ _copy_secrets() {
# run slurmctld
_slurmctld() {
cd /root/rpmbuild/RPMS/$ARCH
yum -y --nogpgcheck localinstall slurm-22.05.6-1.el8.$ARCH.rpm \
slurm-perlapi-22.05.6-1.el8.$ARCH.rpm \
slurm-slurmd-22.05.6-1.el8.$ARCH.rpm \
slurm-torque-22.05.6-1.el8.$ARCH.rpm \
slurm-slurmctld-22.05.6-1.el8.$ARCH.rpm
yum -y --nogpgcheck localinstall slurm-$SLURM_VERSION*.$ARCH.rpm \
slurm-perlapi-$SLURM_VERSION*.$ARCH.rpm \
slurm-slurmd-$SLURM_VERSION*.$ARCH.rpm \
slurm-torque-$SLURM_VERSION*.$ARCH.rpm \
slurm-slurmctld-$SLURM_VERSION*.$ARCH.rpm
echo "checking for slurmdbd.conf"
while [ ! -f /.secret/slurmdbd.conf ]; do
echo -n "."

View File

@ -56,7 +56,7 @@ SchedulerType=sched/backfill
#SchedulerAuth=
#SchedulerPort=
#SchedulerRootFilter=
SelectType=select/cons_res
# SelectType=select/con_res
SelectTypeParameters=CR_CPU_Memory
# FastSchedule=1
#PriorityType=priority/multifactor

View File

@ -1,4 +1,4 @@
FROM clustercockpit/slurm.base:22.05.6
FROM clustercockpit/slurm.base:24.05.3
LABEL org.opencontainers.image.authors="jan.eitzinger@fau.de"
# clean up

View File

@ -3,6 +3,7 @@ set -e
# Determine the system architecture dynamically
ARCH=$(uname -m)
SLURM_VERSION="24.05.3"
SLURM_ACCT_DB_SQL=/slurm_acct_db.sql
@ -52,9 +53,9 @@ _wait_for_worker() {
# run slurmdbd
_slurmdbd() {
cd /root/rpmbuild/RPMS/$ARCH
yum -y --nogpgcheck localinstall slurm-22.05.6-1.el8.$ARCH.rpm \
slurm-perlapi-22.05.6-1.el8.$ARCH.rpm \
slurm-slurmdbd-22.05.6-1.el8.$ARCH.rpm
yum -y --nogpgcheck localinstall slurm-$SLURM_VERSION*.$ARCH.rpm \
slurm-perlapi-$SLURM_VERSION*.$ARCH.rpm \
slurm-slurmdbd-$SLURM_VERSION*.$ARCH.rpm
mkdir -p /var/spool/slurm/d /var/log/slurm /etc/slurm
chown slurm: /var/spool/slurm/d /var/log/slurm
if [[ ! -f /home/config/slurmdbd.conf ]]; then

View File

@ -1,4 +1,4 @@
FROM clustercockpit/slurm.base:22.05.6
FROM clustercockpit/slurm.base:24.05.3
LABEL org.opencontainers.image.authors="jan.eitzinger@fau.de"
# clean up

View File

@ -3,109 +3,73 @@ set -e
# Determine the system architecture dynamically
ARCH=$(uname -m)
SLURM_VERSION="24.05.3"
# start sshd server
_sshd_host() {
if [ ! -d /var/run/sshd ]; then
mkdir /var/run/sshd
ssh-keygen -t rsa -f /etc/ssh/ssh_host_rsa_key -N ''
fi
/usr/sbin/sshd
}
# setup worker ssh to be passwordless
_ssh_worker() {
if [[ ! -d /home/worker ]]; then
mkdir -p /home/worker
chown -R worker:worker /home/worker
if [ ! -d /var/run/sshd ]; then
mkdir /var/run/sshd
ssh-keygen -t rsa -f /etc/ssh/ssh_host_rsa_key -N ''
fi
cat > /home/worker/setup-worker-ssh.sh <<EOF2
mkdir -p ~/.ssh
chmod 0700 ~/.ssh
ssh-keygen -b 2048 -t rsa -f ~/.ssh/id_rsa -q -N "" -C "$(whoami)@$(hostname)-$(date -I)"
cat ~/.ssh/id_rsa.pub > ~/.ssh/authorized_keys
chmod 0640 ~/.ssh/authorized_keys
cat >> ~/.ssh/config <<EOF
Host *
StrictHostKeyChecking no
UserKnownHostsFile /dev/null
LogLevel QUIET
EOF
chmod 0644 ~/.ssh/config
cd ~/
tar -czvf ~/worker-secret.tar.gz .ssh
cd -
EOF2
chmod +x /home/worker/setup-worker-ssh.sh
chown worker: /home/worker/setup-worker-ssh.sh
sudo -u worker /home/worker/setup-worker-ssh.sh
/usr/sbin/sshd
}
# start munge and generate key
_munge_start() {
# start munge using existing key
_munge_start_using_key() {
if [ ! -f /.secret/munge.key ]; then
echo -n "checking for munge.key"
while [ ! -f /.secret/munge.key ]; do
echo -n "."
sleep 1
done
echo ""
fi
cp /.secret/munge.key /etc/munge/munge.key
chown -R munge: /etc/munge /var/lib/munge /var/log/munge /var/run/munge
chmod 0700 /etc/munge
chmod 0711 /var/lib/munge
chmod 0700 /var/log/munge
chmod 0755 /var/run/munge
/sbin/create-munge-key -f
rngd -r /dev/urandom
/usr/sbin/create-munge-key -r -f
sh -c "dd if=/dev/urandom bs=1 count=1024 > /etc/munge/munge.key"
chown munge: /etc/munge/munge.key
chmod 400 /etc/munge/munge.key
sudo -u munge /sbin/munged
munge -n
munge -n | unmunge
remunge
}
# copy secrets to /.secret directory for other nodes
_copy_secrets() {
cp /home/worker/worker-secret.tar.gz /.secret/worker-secret.tar.gz
cp thome/worker/setup-worker-ssh.sh /.secret/setup-worker-ssh.sh
cp /etc/munge/munge.key /.secret/munge.key
rm -f /home/worker/worker-secret.tar.gz
rm -f /home/worker/setup-worker-ssh.sh
}
# run slurmctld
_slurmctld() {
# run slurmrestd
_slurmrestd() {
cd /root/rpmbuild/RPMS/$ARCH
yum -y --nogpgcheck localinstall slurm-22.05.6-1.el8.$ARCH.rpm \
slurm-perlapi-22.05.6-1.el8.$ARCH.rpm \
slurm-slurmd-22.05.6-1.el8.$ARCH.rpm \
slurm-torque-22.05.6-1.el8.$ARCH.rpm \
slurm-slurmctld-22.05.6-1.el8.$ARCH.rpm \
slurm-slurmrestd-22.05.6-1.el8.$ARCH.rpm
yum -y --nogpgcheck localinstall slurm-$SLURM_VERSION*.$ARCH.rpm \
slurm-perlapi-$SLURM_VERSION*.$ARCH.rpm \
slurm-slurmd-$SLURM_VERSION*.$ARCH.rpm \
slurm-torque-$SLURM_VERSION*.$ARCH.rpm \
slurm-slurmctld-$SLURM_VERSION*.$ARCH.rpm \
slurm-slurmrestd-$SLURM_VERSION*.$ARCH.rpm
echo -n "checking for slurmdbd.conf"
while [ ! -f /.secret/slurmdbd.conf ]; do
echo -n "."
sleep 1
done
echo ""
mkdir -p /var/spool/slurm/ctld /var/spool/slurm/d /var/log/slurm /etc/slurm
chown -R slurm: /var/spool/slurm/ctld /var/spool/slurm/d /var/log/slurm
touch /var/log/slurmctld.log
chown slurm: /var/log/slurmctld.log
if [[ ! -f /home/config/slurm.conf ]]; then
# mkdir -p /var/spool/slurm/ctld /var/spool/slurm/d /var/log/slurm /etc/slurm
# chown -R slurm: /var/spool/slurm/ctld /var/spool/slurm/d /var/log/slurm
touch /var/log/slurmrestd.log
chown slurm: /var/log/slurmrestd.log
if [[ ! -f /home/config/slurmrestd.conf ]]; then
echo "### Missing slurm.conf ###"
exit
else
echo "### use provided slurm.conf ###"
cp /home/config/slurm.conf /etc/slurm/slurm.conf
echo "### use provided slurmrestd.conf ###"
cp /home/config/slurmrestd.conf /etc/config/slurmrestd.conf
fi
sacctmgr -i add cluster "snowflake"
sleep 2s
/usr/sbin/slurmctld
cp -f /etc/slurm/slurm.conf /.secret/
/usr/sbin/slurmrestd -f /etc/config/slurmrestd.conf 0.0.0.0:6820 -Dvv
}
### main ###
_sshd_host
_ssh_worker
_munge_start
_copy_secrets
_slurmctld
_munge_start_using_key
_slurmrestd
tail -f /dev/null

View File

@ -0,0 +1,6 @@
#
# Example slurmdbd.conf file.
#
include /etc/slurm/slurm.conf
AuthType=auth/munge

View File

@ -1,4 +1,4 @@
FROM clustercockpit/slurm.base:22.05.6
FROM clustercockpit/slurm.base:24.05.3
LABEL org.opencontainers.image.authors="jan.eitzinger@fau.de"
# clean up

View File

@ -3,6 +3,7 @@ set -e
# Determine the system architecture dynamically
ARCH=$(uname -m)
SLURM_VERSION="24.05.3"
# start sshd server
_sshd_host() {
@ -61,10 +62,10 @@ _start_dbus() {
# run slurmd
_slurmd() {
cd /root/rpmbuild/RPMS/$ARCH
yum -y --nogpgcheck localinstall slurm-22.05.6-1.el8.$ARCH.rpm \
slurm-perlapi-22.05.6-1.el8.$ARCH.rpm \
slurm-slurmd-22.05.6-1.el8.$ARCH.rpm \
slurm-torque-22.05.6-1.el8.$ARCH.rpm
yum -y --nogpgcheck localinstall slurm-$SLURM_VERSION*.$ARCH.rpm \
slurm-perlapi-$SLURM_VERSION*.$ARCH.rpm \
slurm-slurmd-$SLURM_VERSION*.$ARCH.rpm \
slurm-torque-$SLURM_VERSION*.$ARCH.rpm
echo "checking for slurm.conf"
if [ ! -f /.secret/slurm.conf ]; then