Shaheen III

King Abdullah University of Science and Technology · HPE scratch · Fully Reproducible

Files
Fully Reproducible

How the reproducibility levels are defined

System Purpose

Question asked

Please describe the purpose and general usage of the submitted system.  This would include the types of typical applications it supports (e.g., defense applications, molecular dynamics, benchmarking, system test, systems research), and the general use and purpose of the data generated by the applications running on it.

Results obtained on a new HPC system mainly used for general academic reseach as well as AI and machine learning applications.

Availability

Question asked

Please provide the deployment timeframe of the submitted system, or for on-demand cloud systems, the general period over which it is deployed and destroyed.

Please describe the availability of the system to users and who are its set of most regular users.

The system used has been deployed between September and October 2023 and expected to enter actual production before the end of this year.

Exploitation is for a minimum of 5 years and is available for KAUST research but also to other universities and commercial end users.

Storage System Software

Question asked

How is this storage software available? (e.g., commercially, open-source, not publicly available) Note that if the storage software is not open-source or commercially available, then a general description would be requested, but this would limit the submission’s reproducibility score.

Can anyone download/purchase this software?

List either product webpage or open-source repo of the exact software used in IO500.

Give any and all additional details of this specific storage deployment: (e.g., type of storage server product such IBM ESS or DDN SFA400X2, use of Ext4 or some other file system on each storage node, dual connected storage media to storage servers).

The file system is the central parallel file system used by the compute cluster. Data is as diverse as the various field of researches lead by KAUST and their external partnerships and commercial customers.

Software powering the ClusterStor E1000 is under commercial licensing at the exception of the lustre file system itself. The Cray lustre source tree is widely available from github (https://github.com/Cray/lustre)

ClusterStor products description can be found at https://buy.hpe.com/us/en/enterprise-solutions/storage-solutions/cray-clusterstor-storage-systems/cray-clusterstor-e1000-storage-systems/cray-clusterstor-e1000-storage-systems/p/1012842049

This particular file system is a multi-tier storage leveraging both NMVe flash devices for bandwidth and IOPS performances, and HDD providing the bulk of the usable capacity.

Runtime Environment

Question asked

State here that you provided all scripts/documentation that would allow someone else to reproduce your environment and attempt to achieve a similar IO500 score as the submitted system.

NOTE: provide all files/documentation/scripts that would enable a user to build your environment and deploy the custom scripts, software, or config files once they have obtained the appropriate storage system hardware and software. These may be included into the io500.tgz or into the extraCodes upload on the submission form.

Several examples include:

  • Commands used to set striping information (either for the entire system or for particular directories)
  • File system config and tuning information (or a reason why this is not available due to lack of root access, etc) on each node type (e.g., non-default config parameters on all three types of Lustre nodes: MDS, OSS, and client)
  • Any additional scripts utilized that impact IO500 execution beyond the io500 config file. For example, with IBM Spectrum Scale, the output of mmlsconfig, mmlscluster, mmlsfs, and mmlsfileset commands (and possibly even a dump of the configuration if possible). Each file system probably has similar type of config/tuning information that would need to be shared for a user to fully reproduce the environment.

Complete io500 run parmeters can be found in the execution shell script attached to this submission (including striping parameters).

Server side lustre setting are mostly default at the exception of pagecache (disabled for read and writes) and tunings related to the kfilnd LNET driver specific to the SlingShot interconnect.

Client side settings are:

mdc.${fsname}*.max_rpcs_in_flight=128 
mdc.${fsname}*.max_mod_rpcs_in_flight=127 
osc.${fsname}*.max_rpcs_in_flight=256 
osc.${fsname}*.max_dirty_mb=1024 
osc.${fsname}*.max_pages_per_rpc=1024 
osc.${fsname}*.checksums=0 
max_dirty_mb=65536 
llite.${fsname}*.max_read_ahead_mb=2048 
llite.${fsname}*.max_read_ahead_per_file_mb=256 
llite.${fsname}*.max_cached_mb=59392 
ldlm.namespaces.${fsname}*.lru_size=0 

Fault Tolerance Mechanisms

Question asked

Does your system have a single point of failure as defined by “Definition 7” of a Production System? Please describe all mechanisms that provide fault tolerance for the submitted storage system. Be specific to your submission, not general storage system capabilities.

  • Power (For example, can power failure of a node or rack bring down the storage system?)
  • Networking (For example, can failure of a single NIC bring down the storage system? Remember that any networking failure that brings down multiple nodes is acceptable)
  • Inter storage data and metadata server (For example, what type of data protection mechanism does the storage system use to keep operating during node failures, e.g., active-active servers, client-directed RAID, declustered RAID, erasure-coding, 3-way replication)
  • Intra storage data and metadata server storage media (For example, how does the storage eysstem protect against fiailure of a single storage device within a node, e.g., raid5)

Please list any additional information needed to determine whether this system has a single point of failure.

Fully redundant rack power feeds

Multi-rail interconnect with port distribution so any switch can fail without impacting file system services

Dual controller storage with fully redundant components (PSU, IOM, etc.)

Data and metadata protected with GridRAID RAID6 and/or RAID10.

Execution

Question asked

Please provide a description of how the IO500 benchmark was executed, e.g., via system scheduler (e.g., SLURM) to run a job on the compute cluster, which initially ran a setup process to configure the client and file system, and then started the full benchmark.

During the IO500 benchmark execution was the system entirely dedicated to running the benchmark or were there other jobs running in the same cluster and storage system?

Benchmark executed via SLURM:

salloc --comment="IO benchmarking" --exclusive --job-name=IO500 --nodes=2080 --time=0-4:1
./kaust-io500-bandwidth.sh kaust-prod-bandwidth.ini

The entire system was dedicated to running the IO500 benchmark.

Caching

Question asked

Please describe all caching mechanisms in client/server that were utilized during the IO500 run. This could include caching in any storage medium (e.g., SSD, RAM).

A few examples would include:

  • Client data/metadata caching (in Linux page cache or in other memory cache)
  • Client side NVMe read-only data cache
  • Storage server metadata/data caching in RAM
  • Storage controller caching
  • RAID card caching

Client side using pagecache memory (unless DirectIO was used).

Pagecache disabled for read and write on all lustre targets on the servers side

The ClusterStor platform doesn't use any form of writeback cache.

Data Source

Question asked

Is the submitted system a stand-alone storage system or an acceleration layer in front of another storage system that is the source of truth for all application data? This question relates to whether the submitted system is a burst buffer layered on primary storage or primary storage itself.

Standalone file system, no other piece of hardware or software involved

Trust

Question asked

Please describe any steps taken to ensure that the results are trustworthy.

  • Did you run the benchmark multiple times and get similar scores?
  • Did you validate the score is below the physical capabilities of the deployed hardware?
  • How did you validate that the storage system follows the IO500 rules regarding persistence, which states, "All data must be written to persistent storage within the measured time for the individual benchmark, e.g. if a file system caches data, it must ensure that data is persistently stored before acknowledging the close"?

The benchmark was run several times until all results were fully valid. Then, we performed additional runs to ensure results were consistent.

Results obtain match with expected system performances and never exceed theoretical limits (to the extent of measurable hardware and/or software capabilities).

Since the very same file system is used to store binaries, libraries, actual user data and results from this benchmark, we are confident there isn't a single byte of data that's been lost in the process. The platform description and configuration provided should clear doubts regarding any form of caching too.