riverxdata · nttg8100 · Mar 30, 2025 · Mar 30, 2025
diff --git a/blog/2025-01-08-welcome/2025-01-08-index.md b/blog/2025-01-08-welcome/2025-01-08-index.md
@@ -0,0 +1,30 @@
+---
+slug: welcome-to-riverxdata
+title: Welcome to RiverXData
+authors: [river]
+tags: [cloud, data-analysis, slurm, hpc, web-platform]
+---
+
+## Unlock the Power of Cloud-Based Data Analysis with RiverXData
+![RiverXData](./riverxdata.svg)
+
+Welcome to **RiverXData**, a data platform designed for scalable and efficient data analysis. Built on top of **SLURM**, RiverXData brings the power of high-performance computing (HPC) to a user-friendly web-based interface, enabling researchers, engineers, and data scientists to run complex computational tasks with ease.
+
+<!-- truncate -->
+
+### Why RiverXData?
+
+Modern data analysis requires powerful computation, but traditional HPC environments can be challenging to access and manage. RiverXData simplifies this by providing:
+
+- **Seamless Web-Based Access**: Run jobs, manage workloads, and analyze results all from a web browser—no command-line expertise required.
+- **Scalability & Performance**: Harness the power of SLURM for efficient job scheduling and resource management.
+- **Cloud Flexibility**: Deploy, scale, and optimize computational tasks without worrying about infrastructure.
+- **User-Friendly Interface**: Modern web based job in slurm can be accessed interactively
+
+### Get Started Today
+
+Whether you're analyzing large datasets, running simulations, or training machine learning models, RiverXData empowers you to leverage the full potential of cloud-based HPC with minimal setup. 
+
+Stay tuned for tutorials, feature updates, and real-world use cases as we continue to improve RiverXData.
+
+🔗 **Explore more at [RiverXData](#)**
diff --git a/blog/2025-01-08-welcome/riverxdata.svg b/blog/2025-01-08-welcome/riverxdata.svg
diff --git a/blog/2025-03-14/how-to-build-slurm-single-node-with-full-functions.md b/blog/2025-03-14/how-to-build-slurm-single-node-with-full-functions.md
@@ -0,0 +1,356 @@
+---
+slug: single-node-slurm-setup
+title: Single Node Slurm Setup
+authors: [river]
+tags: [slurm, hpc]
+---
+
+
+# Single Node Slurm Setup
+![cluster](./img/Slurm_logo.svg)
+
+SLURM (Simple Linux Utility for Resource Management) is a powerful open-source workload manager commonly used in high-performance computing (HPC) environments. While it is typically deployed across multi-node clusters, setting up SLURM on a single node can be highly beneficial for testing, development, or running lightweight workloads. This guide will help you understand the fundamental concepts of how a scheduler operates and provide a step-by-step walkthrough to configure SLURM with full functionality on a single-node setup.
+
+<!-- truncate -->
+
+**How to Set Up a Fully Functional SLURM Cluster on a Single Node (with Proper Resource Constraints)**
+SLURM is a powerful job scheduler widely used in HPC environments, but configuring it properly—especially on a single-node cluster—can be tricky. Many tutorials cover the basics, like installing SLURM and running jobs, but they often overlook a critical aspect: cgroups (control groups).
+
+Without proper cgroup configuration, SLURM may fail to enforce CPU and memory constraints, allowing jobs to consume more resources than allocated. This can lead to performance degradation, system crashes, or unexpected behavior.
+
+In this guide, I’ll walk you through setting up a fully functional SLURM cluster on a single node, ensuring that CPU and memory limits are properly enforced using cgroups. Whether you're setting up a test environment or a lightweight HPC system, this tutorial will help you avoid common pitfalls and ensure that SLURM effectively manages resources as expected.
+
+Let’s dive in
+
+:::info
+- This architecture is designed for a single node running Ubuntu 20.04.
+- It supports all standard Slurm features.
+- The setup is manual to help you understand how Slurm works. Note that users can utilize resources without submitting jobs, so this configuration is not recommended for production environments.
+:::
+
+## **Install `slurmd` and `slurmctld`**
+
+Install the required software:
+
+```bash
+sudo apt-get update -y && sudo apt-get install -y slurmd slurmctld
+```
+
+Verify the installation:
+
+```bash
+# Locate slurmd and slurmctld
+which slurmd
+# Output: /usr/sbin/slurmd
+which slurmctld
+# Output: /usr/sbin/slurmctld
+```
+
+## **Prepare `slurm.conf`**
+
+:::info
+- This configuration applies to all nodes.
+:::
+
+Create the `slurm.conf` file:
+
+```bash
+cat <<EOF > slurm.conf
+# slurm.conf for a single-node Slurm cluster with accounting
+ClusterName=localcluster
+SlurmctldHost=localhost
+MpiDefault=none
+ProctrackType=proctrack/linuxproc
+ReturnToService=2
+SlurmctldPidFile=/run/slurmctld.pid
+SlurmctldPort=6817
+SlurmdPidFile=/run/slurmd.pid
+SlurmdPort=6818
+SlurmdSpoolDir=/var/lib/slurm-llnl/slurmd
+SlurmUser=slurm
+StateSaveLocation=/var/lib/slurm-llnl/slurmctld
+SwitchType=switch/none
+TaskPlugin=task/none
+
+# TIMERS
+InactiveLimit=0
+KillWait=30
+MinJobAge=300
+SlurmctldTimeout=120
+SlurmdTimeout=300
+Waittime=0
+
+# SCHEDULING
+SchedulerType=sched/backfill
+SelectType=select/cons_tres
+SelectTypeParameters=CR_Core
+
+# ACCOUNTING (slurmdbd, not enabled now)
+AccountingStorageType=accounting_storage/none
+JobAcctGatherType=jobacct_gather/none
+JobAcctGatherFrequency=30
+
+# LOGGING
+SlurmctldDebug=info
+SlurmctldLogFile=/var/log/slurm-llnl/slurmctld.log
+SlurmdDebug=info
+SlurmdLogFile=/var/log/slurm-llnl/slurmd.log
+
+# COMPUTE NODES (Single-node configuration)
+NodeName=localhost CPUs=2 Sockets=1 CoresPerSocket=2 ThreadsPerCore=1 RealMemory=1024 State=UNKNOWN
+
+# PARTITION CONFIGURATION
+PartitionName=LocalQ Nodes=ALL Default=YES MaxTime=INFINITE State=UP
+EOF
+```
+
+Move the file to the correct location:
+
+```bash
+sudo mv slurm.conf /etc/slurm-llnl/slurm.conf
+```
+
+## **Start Basic Slurm Services**
+
+Start the `slurmd` service:
+
+```bash
+# Start the service
+sudo service slurmd start
+# Check its status
+sudo service slurmd status
+```
+
+<figure markdown="span">
+    ![cluster](./img/slurmd_status.png)
+</figure>
+
+Start the `slurmctld` service:
+
+```bash
+# Start the service
+sudo service slurmctld start
+# Check its status
+sudo service slurmctld status
+```
+
+<figure markdown="span">
+    ![cluster](./img/slurmctld_status.png)
+</figure>
+
+Submit a small job (adjust CPUs and memory as needed):
+
+```bash
+srun --mem 500MB -c 1 --pty bash
+# Check details of submitted jobs
+squeue -o "%i %P %u %T %M %l %D %C %m %R %Z %N" | column -t
+```
+
+Before submitting the job, memory usage is less than 200MB:
+
+<figure markdown="span">
+    ![cluster](./img/memory_before_stress.png)
+</figure>
+
+Allocate 100MB of memory repeatedly:
+
+```bash
+declare -a mem
+i=0
+
+while :; do
+    mem[$i]=$(head -c 100M </dev/zero | tr '\000' 'x') 
+    ((i++))
+    echo "Allocated: $((i * 100)) MB"
+done
+```
+
+After allocating 1GB, the job is not killed due to missing control group (cgroup) configuration:
+
+<figure markdown="span">
+    ![cluster](./img/overresource_limit.png)
+</figure>
+
+## **Limit Resources Using cgroups**
+
+Create a `cgroup.conf` file to restrict resource usage:
+
+```bash
+cat <<EOF >cgroup.conf
+CgroupAutomount=yes
+CgroupMountpoint=/sys/fs/cgroup
+ConstrainCores=yes
+ConstrainRAMSpace=yes
+ConstrainDevices=yes
+ConstrainSwapSpace=yes
+MaxSwapPercent=5
+MemorySwappiness=0
+EOF
+```
+
+Move the file to the correct directory:
+
+```bash
+sudo mv cgroup.conf /etc/slurm-llnl/cgroup.conf
+```
+
+Update `slurm.conf` to enable cgroup plugins:
+
+```bash
+sudo sed -i -e "s|ProctrackType=proctrack/linuxproc|ProctrackType=proctrack/cgroup|" \
+            -e "s|TaskPlugin=task/none|TaskPlugin=task/cgroup|" /etc/slurm-llnl/slurm.conf
+```
+
+Enable cgroup in GRUB and reboot:
+
+```bash
+sudo sed -i 's/^GRUB_CMDLINE_LINUX="/GRUB_CMDLINE_LINUX="cgroup_enable=memory swapaccount=1 /' /etc/default/grub
+sudo update-grub
+sudo reboot
+```
+
+Restart Slurm services:
+
+```bash
+sudo service slurmctld restart
+sudo service slurmd restart
+```
+
+Rerun the memory allocation job. This time, the job will be killed when it exceeds the memory limit:
+
+```bash
+srun --mem 500MB -c 1 --pty bash
+declare -a mem
+i=0
+
+while :; do
+    mem[$i]=$(head -c 100M </dev/zero | tr '\000' 'x') 
+    ((i++))
+    echo "Allocated: $((i * 100)) MB"
+done
+```
+
+<figure markdown="span">
+    ![cluster](./img/oom.png)
+</figure>
+
+## **Enable Slurm Accounting**
+
+Accounting allows monitoring of jobs, resource allocation, and permissions.
+
+<figure markdown="span">
+    ![cluster](./img/sacct_disable.png)
+</figure>
+
+### Install `slurmdbd`
+
+```bash
+sudo apt-get install slurmdbd mariadb-server -y
+```
+
+### Configure `slurmdbd.conf`
+
+:::info
+- Enables the accounting plugin to store account information.
+- Maps Linux users to Slurm accounts. Users cannot submit jobs without being added.
+- Useful for monitoring jobs and optimizing resource usage.
+:::
+
+Create the `slurmdbd.conf` file:
+
+```bash
+cat <<EOF >slurmdbd.conf
+PidFile=/run/slurmdbd.pid
+LogFile=/var/log/slurm/slurmdbd.log
+DebugLevel=error
+DbdHost=localhost
+DbdPort=6819
+
+# DB connection data
+StorageType=accounting_storage/mysql
+StorageHost=localhost
+StoragePort=3306
+StorageUser=slurm
+StoragePass=slurm
+StorageLoc=slurm_acct_db
+SlurmUser=slurm
+EOF
+```
+
+Move the file to the correct location:
+
+```bash
+sudo mv slurmdbd.conf /etc/slurm-llnl/slurmdbd.conf
+```
+
+### Create the Database
+
+Create the database and user:
+
+```bash
+sudo service mysql start
+sudo mysql -e "CREATE DATABASE slurm_acct_db;" && \
+sudo mysql -e "CREATE USER 'slurm'@'localhost' IDENTIFIED BY 'slurm';" && \
+sudo mysql -e "GRANT ALL PRIVILEGES ON slurm_acct_db.* TO 'slurm'@'localhost';" && \
+sudo mysql -e "FLUSH PRIVILEGES;"
+```
+
+Verify the database and user:
+
+```bash
+sudo mysql -e "SHOW DATABASES;" 
+sudo mysql -e "SELECT User, Host FROM mysql.user;"
+sudo mysql -e "SHOW GRANTS FOR 'slurm'@'localhost';"
+```
+
+<figure markdown="span">
+    ![cluster](./img/add_db.png)
+</figure>
+
+### Start `slurmdbd` Service
+
+```bash
+sudo service slurmdbd start
+```
+
+Update `slurm.conf` to enable accounting:
+
+```bash
+sudo sed -i -e "s|AccountingStorageType=accounting_storage/none|AccountingStorageType=accounting_storage/slurmdbd\nAccountingStorageEnforce=associations,limits,qos\nAccountingStorageHost=localhost\nAccountingStoragePort=6819|" /etc/slurm-llnl/slurm.conf 
+sudo sed -i -e "s|JobAcctGatherType=jobacct_gather/none|JobAcctGatherType=jobacct_gather/cgroup|" /etc/slurm-llnl/slurm.conf
+sudo systemctl restart slurmctl slurmd
+```
+
+Add Linux users to Slurm accounting:
+
+```bash
+sudo sacctmgr -i add cluster localcluster
+sudo sacctmgr -i --quiet add account $USER Cluster=localcluster
+sudo sacctmgr -i --quiet add user $USER account=$USER DefaultAccount=root
+sudo systemctl restart slurmctl slurmd
+```
+
+<figure markdown="span">
+    ![cluster](./img/add_account.png)
+</figure>
+
+### Submit a Job and View Metrics
+
+Submit a job:
+
+```bash
+srun --mem 500MB -c 1 --pty bash
+```
+
+<figure markdown="span">
+    ![cluster](./img/add_account.png)
+</figure>
+
+## **Conclusion**
+
+:::info
+- Slurm is widely used in academic and industrial settings for orchestrating distributed jobs across multiple nodes.
+- While Slurm is relatively easy to set up, critical steps like resource limits and accounting are often overlooked.
+- Slurm integrates seamlessly with distributed computing frameworks like Spark, Ray, Dask, and Flink, enabling efficient resource utilization for local development.
+:::
+
diff --git a/blog/2025-03-14/img/HPC_architecture.png b/blog/2025-03-14/img/HPC_architecture.png
diff --git a/blog/2025-03-14/img/NFS.png b/blog/2025-03-14/img/NFS.png
diff --git a/blog/2025-03-14/img/Slurm_logo.svg b/blog/2025-03-14/img/Slurm_logo.svg
diff --git a/blog/2025-03-14/img/account_usage.png b/blog/2025-03-14/img/account_usage.png
diff --git a/blog/2025-03-14/img/add_account.png b/blog/2025-03-14/img/add_account.png
diff --git a/blog/2025-03-14/img/add_db.png b/blog/2025-03-14/img/add_db.png
diff --git a/blog/2025-03-14/img/config_1.png b/blog/2025-03-14/img/config_1.png
diff --git a/blog/2025-03-14/img/config_2.png b/blog/2025-03-14/img/config_2.png
diff --git a/blog/2025-03-14/img/config_3.png b/blog/2025-03-14/img/config_3.png
diff --git a/blog/2025-03-14/img/create_app.png b/blog/2025-03-14/img/create_app.png
diff --git a/blog/2025-03-14/img/grafana.png b/blog/2025-03-14/img/grafana.png
diff --git a/blog/2025-03-14/img/grafana_down.png b/blog/2025-03-14/img/grafana_down.png
diff --git a/blog/2025-03-14/img/job_sinfo.png b/blog/2025-03-14/img/job_sinfo.png
diff --git a/blog/2025-03-14/img/memory_before_stress.png b/blog/2025-03-14/img/memory_before_stress.png
diff --git a/blog/2025-03-14/img/node_grafana.png b/blog/2025-03-14/img/node_grafana.png
diff --git a/blog/2025-03-14/img/oom.png b/blog/2025-03-14/img/oom.png
diff --git a/blog/2025-03-14/img/overresource_limit.png b/blog/2025-03-14/img/overresource_limit.png
diff --git a/blog/2025-03-14/img/resume_node.png b/blog/2025-03-14/img/resume_node.png
diff --git a/blog/2025-03-14/img/sacct_disable.png b/blog/2025-03-14/img/sacct_disable.png
diff --git a/blog/2025-03-14/img/slurm.svg b/blog/2025-03-14/img/slurm.svg
diff --git a/blog/2025-03-14/img/slurm_arch.gif b/blog/2025-03-14/img/slurm_arch.gif
diff --git a/blog/2025-03-14/img/slurm_grafana.png b/blog/2025-03-14/img/slurm_grafana.png
diff --git a/blog/2025-03-14/img/slurmctld_status.png b/blog/2025-03-14/img/slurmctld_status.png
diff --git a/blog/2025-03-14/img/slurmd_status.png b/blog/2025-03-14/img/slurmd_status.png
diff --git a/blog/2025-03-14/img/small_HPC.jpg b/blog/2025-03-14/img/small_HPC.jpg
diff --git a/blog/2025-03-14/img/srun_fastqc.png b/blog/2025-03-14/img/srun_fastqc.png
diff --git a/blog/2025-03-14/img/submit_job.png b/blog/2025-03-14/img/submit_job.png
diff --git a/blog/2025-03-28/1k_threads.png b/blog/2025-03-28/1k_threads.png
diff --git a/blog/2025-03-28/cpu_thread.png b/blog/2025-03-28/cpu_thread.png
diff --git a/blog/2025-03-28/process_by_htop.png b/blog/2025-03-28/process_by_htop.png
diff --git a/blog/2025-03-28/simple.png b/blog/2025-03-28/simple.png