v0.7.0
Highlights
- SLURM-like experience
- Storage support (GCSFuse, GCP Filestore)
- Remote state management for Cluster Toolkit backend
What's Changed
New Features
- Support for gcs_bucket added to blueprint_generator by @sharabiani in #351
- Add
xpk runcommand by @IrvingMg in #343 - Add installation of gke-gcloud-auth-plugin by @IrvingMg in #357
- Implement config commands by @pawloch00 in #354
- Merge gcsfue into main by @pawloch00 in #317
- Merge gcsfuse into develop by @pawloch00 in #370
- Implement gcs state bucket parameter passing by @pawloch00 in #377
- Implement remote state managment changes by @pawloch00 in #380
- Add environment variables for Pathways metric collection by @guptaaka in #404
- Kjob storage configuration by @mbobrovskyi in #372
- Kjob storage configuration for run command by @mbobrovskyi in #408
- Support for multihost in slurm mode by @pawloch00 in #409
xpk storage detachby @gcie in #420
Bug fixes
- Fix flaky
Check created jobtest by @IrvingMg in #361 - Fix
xpk versionshowing invalid git commit hash by @gcie in #358 - Sync main commits with develop branch by @SujeethJinesh in #379
- Add Workaround for Megascale Num Slices Issue on PW by @SujeethJinesh in #375
- Fix invalid nvidia libraries mount path for A3 Mega by @gcie in #382
- Skip Storages if Not Found (instead of failing) by @SujeethJinesh in #381
- De-dupe Superfluous Exit Codes Provided by the User by @SujeethJinesh in #387
- Update Resource Manager Kueue CPU Requirements by @SujeethJinesh in #386
- Add Host Networking with Cluster First DNS to Main Job by @SujeethJinesh in #391
- Fixes for slurm-like commands by @sharabiani in #373
- Fixes for slurm like commands by @pawloch00 in #397
- Fix nightly tests by @mbobrovskyi in #415
- Sync develop with main by @sharabiani in #421
- Fix xpk version error when cloned from git by @sharabiani in #423
New Contributors
Full Changelog: v0.6.0...v0.7.0