Release v0.7.5

In this release I one big problem was identified and solved - it's API performance, but you need to regenerate hashes for the users passwords to make it work properly. Also it contains the quality of life changes to ease day-to-day migration procedures. Last one is CI now validating linting and some benchmarks, so as usual all the files was changed to please the rules.

What's Changed

• Fix for hardcoded image wait timeout on TaskImage by @sparshev in #76
• API Perf improvement and Fish cpu/mem limitations by @sparshev in #77
• Allow user to change its password by @sparshev in #83
• Added golangci-lint and multiple fixes for all kind of checks by @sparshev in #84

Full Changelog: v0.7.4...v0.7.5

Release v0.7.4

What's Changed

• Add initial SSH Proxy for VMX by @svenevs in #72
• Add TaskImage for AWS and an ability to use asterisk in image by @sparshev in #74
  • Encryption (re-encryption) - so even if you using not encrypted images as a base - the produced one will be encrypted
  • Latest image by the name with asterisk - actually AWS feature, but now logic tries to find the latest image if there is multiple of them in the return.

Other changes:

• Fixed a couple of issues with openapi spec that failed the jenkins plugin build and added openapi spec yaml validation to CI.
• Switched to the newest version of openapi-codegen. Also tried ogen and it could solve some issues with code generation, but needs a major rewrite to use it (#75).
• Added new API call to get tasks by ID - that helps to wait and find out how the task completed and what's the image name/id we've got.
• Added more logging to tasks to keep track of their execution.
• Made sure NodeFilter field in Resources will not return null and just empty list as described in spec.
• Small fix to not print timestamps in debug mode if we don't need timestamps.

Full Changelog: v0.7.3...v0.7.4

Release v0.7.3

What's Changed

• CI: In dash interp jobs -r is not supported, so cutting it off by @sparshev in #58
• CI: Add patch version number in go.mod (X.Y => X.Y.Z) by @svenevs in #68
• Common: Added pprof into the node API to debug memory leaks by @sparshev in #67
• Common: Multi-driver configs and the other useful fixes by @sparshev in #70
• AWS: Dedicated hosts management by @sparshev in #66

New Contributors

• @svenevs made their first contribution in #68

Full Changelog: v0.7.2...v0.7.3

Release v0.7.2

Bugfix release for Native driver and some improvements for the build & test system.

What's Changed

• Native driver bugfixes and org improvements by @sparshev in #57

Full Changelog: v0.7.1...v0.7.2

Release v0.7.1

• Fixed a small issue during running an actual mac machine using AWS provider.
• Also btw fixed sed command during the build on mac, because was incompatible with BSD sed

Update 11/10/23: Reuploaded the linux artifacts due to built on 22.04 - and need to be compatible with lts-1 at least (20.04). Will fix release process in the next change.

Release v0.7.0

New native driver, yaml interface and more bugfixes!

What's Changed

• YAML support, Maintenance/Shutdown API and updated dependencies by @sparshev in #49

  This change is for improving the general usability:

  • Added yaml support to all API requests to improve user quality of life - storing labels in YAML is much easier than json, json interface is still supported
  • Added maintenance / shutdown API - to allow user to request graceful shutdown of the node, will be really useful in the loaded cluster #30
  • Updated the dependencies - there was a number of updated deps which needed the updates
  • Cleaned up the logic a bit

• Added native driver and redesigned the image management logic by @sparshev in #46

  Besides adding the new Native driver for MacOS this change also changes the image definition process of the VMX and Docker drivers.

  • Native driver was added! It's as usual supports only macos for now, but adding to the other systems should not be so hard. It uses the node description to get the identifiers (configurable) which should be used by the Labels to find the appropriate fish node to execute on.
  • BREAKING: VMX and Docker now uses a bit different way to specify the Label definition images, so the Labels need to be updated. But as an advantage now it supports checksum to verify the image!
  • Added node filter and node identifiers for easy selecting where to execute the workload. It's common for all the drivers so can be used for VMX to run macos on macos or to run arm64 on arm64.
  • Switched examples to YAML representation - now it's readability is better
  • A number of fixes:
    • Fixed AWS startup issue - messed up gt with lt as usual
    • Replaced ioutil usage since deprecated as of go v1.16
    • Replaced all interface{} to any
    • Fixed YAML user interface introduced in #48 - now it properly handled
    • Added UID references validation, previously they was handled automatically but after migration from dqlite that was lost

Full Changelog: v0.6.4...v0.7.0

Release v0.6.4

Bugfix release: A nasty bug was in fish for quite a long time, affected the restart of the node with allocated applications and they were unable to continue operaions and when VMX machines are not stopped correctly through soft stop.

What's Changed

• BUGFIX: The changes about node resources monitoring and drivers placement messed up the node restart procedure if there was allocated resources, so fixed it and added integration test to fight the potential regression by @sparshev in #47
• BUGFIX: Fixed the issue with not stopping VMX VM's if they don't want to do that through soft procedure by @sparshev in #47

Full Changelog: v0.6.3...v0.6.4

Release v0.6.3

Bugfix release

What's Changed

• Cleanup is quite an important feature, otherwise automated Fish is clogging the disk or leftovers from the bad allocations. Also it's good to clean if images was not downloaded/unpacked correctly (I'm looking at you, max opened files on mac!) by @sparshev in #45
• Removing driver from the list if it failed preparation phase - if they are not prepared, so what's the point of having them in the roster... by @sparshev in #45
• 1 minute retry AWS connection on start in case network is not working - caught on mac after reboot, service starts and gets no internet... by @sparshev in #45
• Fixed nasty issue with waiting of AWS instance IP and disks - it sometimes don't get the IP immediately and just waits for 5 microseconds and repeats with 0 result of course by @sparshev in #45

Full Changelog: v0.6.2...v0.6.3

Release v0.6.2

A couple of fixes for AWS and CI system

What's Changed

• Small fixes and doc by @sparshev in #43
• AWS: Fixed the issue with incorrect AWS quotas calculation and moved docs to wiki by @sparshev in #44

Full Changelog: v0.6.1...v0.6.2

Release v0.6.1

First automatic release with a couple of fixes and multiple supported platforms.