Gyubong Lee
Gyubong Lee
Partially fix #2337. # Test Manually tested it using the following methods. In this PR, the `cr.backend.ai/stable/python` image is used as a placeholder for testing. ## Prerequisite 1. Tag and...
Fix #2359, ref #1683. **Checklist:** (if applicable) - [x] Milestone metadata specifying the target backport version - [x] Mention to the original issue
Partially fix #2337. WIP **Checklist:** (if applicable) - [x] Milestone metadata specifying the target backport version - [x] Mention to the original issue
Resolves https://github.com/lablup/giftbox/issues/638. Implement API for querying GPU allocation map (GPU allocation states per GPU device). The GPU allocation is calculated by reading the *resource.txt* file in the scratch directory per...
Partially fix #1634. **Checklist:** (if applicable) - [x] Milestone metadata specifying the target backport version - [x] Mention to the original issue
Partially fix #1634. **Checklist:** (if applicable) - [x] Milestone metadata specifying the target backport version - [x] Mention to the original issue
Partially fix #1634. **Checklist:** (if applicable) - [x] Milestone metadata specifying the target backport version - [x] Mention to the original issue
Since `DeviceId` is a *str* type, [mother_uuid](https://github.com/lablup/backend.ai/blob/14996f2c8ea13301a6d57b1a7baa2e7dc7512b93/src/ai/backend/accelerator/mock/plugin.py#L109) should be defined as `t.String`. https://github.com/lablup/backend.ai/blob/14996f2c8ea13301a6d57b1a7baa2e7dc7512b93/src/ai/backend/accelerator/mock/plugin.py#L104-L118 Currently, because `mother_uuid` is defined as `tx.UUID`, the following bug is occurring unintentionally when trying to creating...
This PR aims to resolve the distribution locking issues by integrating [Raftify](https://github.com/lablup/raftify) with Backend.AI manager (based on GlobalTimer operating by Raft algorithm). Any kind of feedback is welcome. > Note:...
This PR changes the existing multi-node, multi-process design in the Raftify integration PR (#2105). Specifically, it modifies the design from creating num-proc RaftNodes per physical node to having only one...