v1.7.0: Improved blueprint validation, official support for Ubuntu, and bug fixes for Slurm v5
Key New Features
- A new test_apis_enabled validator will check whether Google Cloud services required by the blueprint are enabled
- The Cloud HPC Toolkit has added official support for Ubuntu 20.04 images
Improvements
- Batch modules now support Shared VPC networks
- VPC module enables jumbo frames by default
- AMD-optimized blueprint includes the OpenFOAM application compiled for the Zen 3 microarchitecture
- A new example blueprint demonstrates using local SSD disks with VM instances
- A new example blueprint demonstrates installation of StarCCM+ CFD simulation application
Bug Fixes
- Resolve Slurm v5 startup-script timeout errors
by demonstrating the use of a build VM to install Spack and optimized applications - Fix incompatibility between DDN Exascaler (Lustre) and Slurm v5 modules
What's Changed
- Add shared task for recover from terraform failure by @nick-stroud in #578
- Automatically pull serial port 1 logs on startup script failure by @nick-stroud in #580
- Bump github.com/hashicorp/hcl/v2 from 2.14.0 to 2.14.1 by @dependabot in #572
- Enable shared VPC to be used with Batch modules by @nick-stroud in #573
- Update default MTU for new VPCs by @tpdownes in #577
- Replace HTCondor integration test by @tpdownes in #583
- Remove blueprint_dir update in integration tests by @heyealex in #582
- Add documentation that cloud-batch-job supports shared vpc by @nick-stroud in #584
- Add Spack setup runner by @nick-stroud in #586
- Implement API validation by @tpdownes in #576
- Add documentation for API validation by @tpdownes in #585
- Document troubleshooting for Slurm startup script timeouts by @tpdownes in #590
- Bump github.com/zclconf/go-cty from 1.10.0 to 1.11.0 by @dependabot in #491
- Update wait for startup for qsim example to 2400s by @heyealex in #591
- Bump google.golang.org/api from 0.96.0 to 0.98.0 by @dependabot in #589
- Bump github.com/googleapis/gax-go/v2 from 2.4.0 to 2.5.1 by @dependabot in #588
- Fix test_apis_enabled to handle empty API list by @tpdownes in #592
- Add default Spack log file by @nick-stroud in #595
- Fix: passing non-existent cache was causing failures in some projects by @nick-stroud in #593
- Batch modules should require Batch API by @tpdownes in #597
- Support local module required APIs by @tpdownes in #596
- hybrid-slurm documentation for on-premise by @heyealex in #579
- Add ability to pass non-text variables through the cli
--vars
flag by @kkr16 in #574 - Additional Batch parameters by @nick-stroud in #594
- Add slurm v5 HPC Centos example to integration tests by @heyealex in #587
- Clarify and compact validator failure message by @tpdownes in #598
- Add slurm-gcp-v5-ubuntu2004 example to integration tests by @heyealex in #601
- Remove deployment_name var from vpc example as it is passed implicitly by @nick-stroud in #605
- Add variable to turn off public ips on Batch jobs by @nick-stroud in #609
- Update AMD blueprint with OpenFOAM application and test suite by @tpdownes in #606
- Filestore cleanup fails on existing instances by @heyealex in #610
- Develop by @omartin2010 in #608
- Add image support documentation by @heyealex in #611
- Push test deployment directory to GCS by @heyealex in #602
- release-candidate - Handle case that cloud-batch-job does not have startup script in metadata by @nick-stroud in #612
- Add a boilerplate blueprint to README by @nick-stroud in #617
- Back-porting EXAScaler fix for V5 on release candidate by @cboneti in #624
- Release Candidate: Simcenter StarCCM+ Tutorial Blueprint by @heyealex in #637
- Resolve conflicts on release-candidate by @nick-stroud in #646
- Release v1.7.0 by @tpdownes in #649
New Contributors
- @omartin2010 made their first contribution in #608
Full Changelog: v1.6.0...v1.7.0