Advanced Training:

Data to Compute

May 28-31, 2024

iRODS User Group Meeting 2024

Amsterdam, Netherlands

Alan King, Senior Software Developer

Martin Flores, Software Developer

iRODS Consortium

Integrating iRODS with a compute environment

In order of increasing complexity and integration...

 

iRODS as a compute orchestrator

  • Launch a job via irule, or as part of a PEP
  • Implement a Landing Zone for product capture

 

iRODS as part of a compute job script    (TODAY)

  • Stage the source data via replication for the application
  • Capture the products and ingest them into iRODS

 

iRODS as part of the compute application

  • Compute application directly leverages the iRODS API to open, read, and write data

The Data to Compute Use Case

Focus on the right side of the picture

iRODS is out of the data path for HPC computation

Goal - Develop generic interface concept for compute

  • Develop a metadata-driven interface to derive path for input data and compute results. Utilize it to:
    • push data to the proper storage resource
    • get a name for the host on which to launch compute job(s)

 

  • Separate configuration from implementation
    • Keep deployment specifics in configuration files
    • Keep rulebase, scripts, and modules free of hard-wired values

Goal - Develop a thumbnailing service for iRODS

Interface is through iRODS and SLURM (compute job scheduler):

  • Replicate the data to the compute resource
  • Send a job to the compute scheduler to generate thumbnails
  • Register the thumbnails into the catalog
  • Replicate the thumbnails back to long term storage
  • Trim replicas on compute resource

Components of the System

System Component

Job Scheduler

Job Launching Script

Tools to Execute

Job Endpoint

Implementation

SLURM

bash

Image Magick convert

iRODS Rule Base

(user extension of the iRODS API)

and SLURM prolog / epilog

Getting Started

sudo apt-get update

sudo apt-get -y install python3-pip irods-rule-engine-plugin-python imagemagick

 

sudo su - irods

Install the PRC (Python iRODS-Client) module for irods.

As the ubuntu user ...

python3 -m pip install --upgrade pip
python3 -m pip install python-irodsclient

Install Image Magick, pip, and Python Rule Engine Plugin and switch to irods user.

Getting Started

As the ubuntu user... Get the irods_training repository

cd
git clone https://github.com/irods/irods_training
sudo apt-get -y install irods-externals-cmake3.21.4-0 irods-dev
export PATH=/opt/irods-externals/cmake3.21.4-0/bin:$PATH
cd
mkdir build_data_to_compute
cd build_data_to_compute
cmake ../irods_training/advanced/hpc_data_to_compute/
make package
sudo dpkg -i irods-hpc-data-to-compute-example.deb

Install, configure, and start SLURM (job scheduler)

Build and Install the Data-to-Compute package

cd ~/irods_training/advanced/hpc_data_to_compute/
./ubuntu_22/install_and_configure_slurm.sh
sudo systemctl restart slurmd slurmctld

Package Contents

$ dpkg -c irods-hpc-data-to-compute-example.deb
drwxrwxr-x root/root         0 2024-05-12 00:45 ./etc/
drwxrwxr-x root/root         0 2024-05-12 00:45 ./etc/irods/
-r--r--r-- root/root      1213 2024-05-12 00:40 ./etc/irods/core.py.data_to_compute
-r--r--r-- root/root      1144 2024-05-12 00:40 ./etc/irods/data_to_compute.re
drwxrwxr-x root/root         0 2024-05-12 00:45 ./var/
drwxrwxr-x root/root         0 2024-05-12 00:45 ./var/lib/
drwxrwxr-x root/root         0 2024-05-12 00:45 ./var/lib/irods/
drwxrwxr-x root/root         0 2024-05-12 00:45 ./var/lib/irods/compute/
-r--r--r-- root/root         0 2024-05-12 00:40 ./var/lib/irods/compute/__init__.py
-rw------- root/root        59 2024-05-12 00:40 ./var/lib/irods/compute/admin_as_rodsuser.json
-r--r--r-- root/root     11773 2024-05-12 00:40 ./var/lib/irods/compute/common.py
-r--r--r-- root/root      2005 2024-05-12 00:40 ./var/lib/irods/compute/irods_compute_functions
-r--r--r-- root/root       565 2024-05-12 00:40 ./var/lib/irods/compute/job_params.json
-r--r--r-- root/root      2150 2024-05-12 00:40 ./var/lib/irods/compute/util.py
-r--r--r-- root/root      1302 2024-05-12 00:40 ./var/lib/irods/detect_thumbnails.py
drwxrwxr-x root/root         0 2024-05-12 00:45 ./var/lib/irods/msiExecCmd_bin/
-r-xr-xr-x root/root       383 2024-05-12 00:40 ./var/lib/irods/msiExecCmd_bin/convert.SLURM
-r-xr-xr-x root/root       328 2024-05-12 00:40 ./var/lib/irods/msiExecCmd_bin/submit_thumbnail_job.sh
-r--r--r-- root/root       788 2024-05-12 00:40 ./var/lib/irods/rescName_from_kvpair.r
-r--r--r-- root/root      1494 2024-05-12 00:40 ./var/lib/irods/spawn_remote_slurm_jobs.r

Configure the rule engine

As the irods user, add an additional rule base to

 /etc/irods/server_config.json :
"rule_engines": [

    ...

        "re_rulebase_set": [

            "data_to_compute",

            "core"

        ],

    ...

]

Add in a small python rulebase:

As the irods user,

create an /etc/irods/core.py with the following content:

def pyParseRoleSpec (rule_args,callback,rei):
    compute_resc_spec = rule_args[0]
    kvp = compute_resc_spec.split('=')+['']
    rule_args[1:3] = list(map(lambda x: x.strip(), kvp[:2]))

Python Rule Engine Configuration

Insert the python plugin configuration stanza into /etc/irods/server_config.json:

"rule_engines": [
    {
         "instance_name": "irods_rule_engine_plugin-irods_rule_language-instance",

                   ...

    },
    {
         "instance_name" : "irods_rule_engine_plugin-python-instance",
         "plugin_name" : "irods_rule_engine_plugin-python",
         "plugin_specific_configuration" : {}
    }, 
​​

Configure the LTS and Image Processing Resources

As the irods user:

 

Make two unix file system resources

iadmin mkresc lts_resc unixfilesystem $(hostname):/tmp/irods/lts_resc
iadmin mkresc img_resc unixfilesystem $(hostname):/tmp/irods/img_resc

Annotate them with appropriate metadata given their roles

  - defined in the configuration as part of the contract

imeta add -R lts_resc COMPUTE_RESOURCE_ROLE LONG_TERM_STORAGE
imeta add -R img_resc COMPUTE_RESOURCE_ROLE IMAGE_PROCESSING
cp ~/irods_training/stickers.jpg /tmp
sudo mkdir -p /tmp/irods/thumbnails
sudo chown -R irods:irods /tmp/irods

As the ubuntu user:

 

Stage data and destination directory for thumbnail creation

The configuration interface

Define interfaces for any necessary conventions

  • Metadata attributes and values

  • Metadata values for implemented roles

  • Interface to job scheduler for launching compute

 

Single Point of Truth - allows for the use of the same 'end-points' for various metadata standards and naming conventions

 

Users may utilize metadata conventions to provide inputs to a given compute job

The configuration interface

For the thumbnail service we will need to

  • Get the metadata attribute string that holds the role

  • Get the tag for an Image Compute resource

  • Get the tag for a Long Term Storage resource

  • Get the logical collection name for thumbnails

  • Get the physical path for a thumbnail

  • Get the name of a thumbnail

  • Get a list of desired thumbnail sizes

The configuration interface

iRODS rule file provides interface for job submission

testRule
{
  *thumbnail_sizes = "128x128,256x256,512x512,1024x1024"
  *host = ""
  *key = "COMPUTE_RESOURCE_ROLE"; *val="IMAGE_PROCESSING"
  # or even: # *key = "COMPUTE_RESOURCE_ROLE=IMAGE_PROCESSING"; *val=""
  *resc_name = ""
  get_host_and_resource_name_by_role(*host, *resc_name, *key, *val)
  writeLine ("stdout", "host=[*host] resc=[*resc_name]")
  writeLine ("stdout", "thumbnails to generate : [ *thumbnail_sizes ]")
  *input_file_name = "stickers"
  *input_file_ext = ".jpg"
  *input_file =  "*input_file_name" ++ "*input_file_ext"
  if ("*host" == "" ) {
    writeLine ("stdout", "Host for job launch was not found.")
  } else {
    foreach (*x in select DATA_PATH where COLL_NAME = '/tempZone/home/rods' and
                   DATA_NAME = '*input_file' and RESC_NAME = '*resc_name')
    {
     *src_phy_path = *x.DATA_PATH
    }
    remote(*host, "") {
      writeLine("serverLog", "-----> remoteExec on host [*host]:")
      foreach (*size in split (*thumbnail_sizes, ",")) {
        *dst_phy_path = "/tmp/irods/thumbnails/" ++ "*input_file_name" ++ "_thumbnail_" ++ "*size" ++ "*input_file_ext"
        writeLine("serverLog"," - thumbsize [ *size ]; convert ( *src_phy_path , *dst_phy_path )")
        *cmd_opts="/var/lib/irods/msiExecCmd_bin/convert.SLURM -thumbnail *size *src_phy_path *dst_phy_path"
        msiExecCmd("submit_thumbnail_job.sh","*cmd_opts","null","null","null",*OUT)
      } #foreach
    } #remote
  } #if-else
} # end rule

INPUT  null
OUTPUT ruleExecOut

The configuration interface

Abstraction of job submission via shell script

#!/bin/bash

# $1 - executable
# $2 - thumbnail option
# $3 - sizing string
# $4 - source physical path
# $5 - destination physical path

SBATCH_OPTIONS="-o /tmp/slurm-%j.out"

SCRIPT="$1" # assume full path to executable

/usr/local/bin/sbatch $SBATCH_OPTIONS "$SCRIPT" \
    ${2+"$2"} \
    ${3+"$3"} \
    ${4+"$4"} \
    ${5+"$5"} \
    >/dev/null 2>&1

 

Thumbnail Service - testing

$ COMPUTE_RESC=$(iquest %s " select RESC_NAME where META_RESC_ATTR_NAME = \
'COMPUTE_RESOURCE_ROLE' and META_RESC_ATTR_VALUE = 'IMAGE_PROCESSING' " )
$ iput -R ${COMPUTE_RESC} /tmp/stickers.jpg
$ ils -l
/tempZone/home/rods:
  rods              0 img_resc      2157087 2024-05-05.22:39 & stickers.jpg
$ imkdir stickers_thumbnails
$ irule -r irods_rule_engine_plugin-irods_rule_language-instance -F spawn_remote_slurm_jobs.r
$ squeue
             JOBID PARTITION     NAME     USER ST       TIME  NODES NODELIST(REASON)
                 4     debug convert.    irods PD       0:00      1 (Resources)
                 5     debug convert.    irods PD       0:00      1 (Priority)
                 2     debug convert.    irods  R       0:01      1 ip-172-31-44-38
                 3     debug convert.    irods  R       0:01      1 ip-172-31-44-38
$ squeue 
             JOBID PARTITION     NAME     USER ST       TIME  NODES NODELIST(REASON)
$ iquest "%s : %s/%s" "select RESC_NAME,COLL_NAME,DATA_NAME where DATA_NAME like '%thumbnail%'"
lts_resc : /tempZone/home/rods/stickers_thumbnails/stickers_thumbnail_1024x1024.jpg
lts_resc : /tempZone/home/rods/stickers_thumbnails/stickers_thumbnail_128x128.jpg
lts_resc : /tempZone/home/rods/stickers_thumbnails/stickers_thumbnail_256x256.jpg
lts_resc : /tempZone/home/rods/stickers_thumbnails/stickers_thumbnail_512x512.jpg

 

(Wait for the SLURM job queue to be empty: )

As the irods user, position the input data and start the thumbnail jobs:

Extending iRODS with the Rule Engine

  • All rules should be created and tested in user space before being installed as a rule base

  • Rules may be refactored into a microservice plugin

  • Rules may be refactored into a C++ rule engine plugin

  • Rules may be refactored into an API plugin

Questions?