Tag: v1.1.4
Commit: 1f49d8d
Notable change:
- Support any region on AWS, and removed requirement of AMIs. The AMI is not being used by default.
- Container import & export: Directly use container images with other orchestration engine
- mTLS support for validating agent identity
What's Changed
- Add more progress logging for instance creation by @eagleonhill in #212
- Fix server stuck terminating if a session is scheduling by @eagleonhill in #213
- Add command to export current instance as a container image by @eagleonhill in #214
- Separate workload and sandbox daemon cgroup for resource monitoring by @eagleonhill in #215
- Set TCP_NOTSENT_LOWAT for priority to work by @eagleonhill in #209
- File server use same auth method as NFS by @eagleonhill in #216
- Separate agent & controller image build pipeline, add fileserver to r… by @eagleonhill in #217
- Fix name by @eagleonhill in #218
- remove download step in packer by @eagleonhill in #219
- Add document for CAS by @eagleonhill in #220
- Make container export output consistent by @eagleonhill in #222
- Add exclude option & strip file timestamps by @eagleonhill in #223
- Add automatic layering based on directory size by @eagleonhill in #224
- Use a streaming grpc to receive task log message. fix some integratio… by @eagleonhill in #225
- Remove support for task log on instance-storage by @eagleonhill in #226
- Add user agent to request by @eagleonhill in #227
- Support more nebius disk type and shared file-system by @eagleonhill in #228
- Add option for non-priveleged agents by @eagleonhill in #229
- Add merkle tree based hashing for layer digest by @eagleonhill in #230
- Move NFS auth to storage interface, implement as common LocalNfsAuth … by @eagleonhill in #232
- Add keep-alive support for storage by @eagleonhill in #233
- Add region to create instance by @eagleonhill in #234
- add volume binding by @eagleonhill in #235
- Add validation for volume name by @eagleonhill in #236
- Support system-ami mode in aws, agent version override. by @eagleonhill in #237
- Report & handle error of scaling error, pool can report last error by @eagleonhill in #238
- Fail immediately in vbatch if pool is unknown by @eagleonhill in #239
- Reserve limit workload memory to avoid crash the entire system by @eagleonhill in #240
- Move the init-from-docker to be fully done on remote through a batch job. by @eagleonhill in #241
- Remove hard-coded UID by @eagleonhill in #242
- Add size status report in import, make it init script more portable by @eagleonhill in #243
- Allow instance to be kept in db for terminated status, by @eagleonhill in #244
- Log streaming use shared channel to avoid race by @eagleonhill in #245
- Fix quota not returned on renew failure by @eagleonhill in #246
- Fix some race condition on follow logs by @eagleonhill in #248
- Refactor completion handler to use typed interface; by @eagleonhill in #247
- Switch nebius to use stopped instance for parking by @eagleonhill in #249
- Improve behavior on suspending by @eagleonhill in #250
- Make cgroup subtree_control non-failing by @eagleonhill in #251
- Kill all sessions when deleting instances by @eagleonhill in #252
- Enable encryption in nebius disks by default by @eagleonhill in #253
- Default to enable dns search in instance creation by @eagleonhill in #254
- Revoke file access on connection drop by @eagleonhill in #255
- Use domain socket for per-agent debug by @eagleonhill in #256
- Add mTLS support for agent to verify agent identity by @eagleonhill in #257
- New worker marked as pending if currently deleting by @eagleonhill in #258
- E2E testing for AWS instance without AMI by @eagleonhill in #259
- Remove AMI build in release pipeline by @eagleonhill in #260
Full Changelog: v1.1.3...v1.1.4