Skip to content
This repository has been archived by the owner on Oct 9, 2023. It is now read-only.

Releases: flyteorg/flyteplugins

Add missing log links for AWS Batch

20 Mar 23:28
fd493e3
Compare
Choose a tag to compare
Add missing sub-task log links (#62)

* Add missing sub-task log links

* PR Comments

* Update deps

* Only update stdlib

* Unit tests

Fetch whether task is interruptible from taskMetaData

19 Mar 16:35
798212c
Compare
Choose a tag to compare

This flag will enable usage of spot instances in k8s clusters.

Replace lower-cased sidecar gpu resource with nvidia one

09 Mar 18:58
e79dce8
Compare
Choose a tag to compare
v0.3.11

Replace lower-cased sidecar gpu resource with nvidia one (#65)

Spark plugin: Scala Support

04 Mar 21:51
95f56f3
Compare
Choose a tag to compare
Merge pull request #63 from lyft/spark-scala-class

Bug: Scala Requires Both MainApplicationFile and MainClass to be Set

Fairness V0 in resource manager: adding support for project-level and namespace-level cap

03 Mar 22:11
f3a131a
Compare
Choose a tag to compare
Merge pull request #56 from lyft/fairness-namespace-cap

Fairness in resource manager: adding support for project-level and namespace-level cap

bumping up the flyteidl version

29 Feb 00:06
9ac5bea
Compare
Choose a tag to compare
v0.3.8

bumping up the flyteidl version (#61)

Distinguish system failures from user

28 Feb 18:59
28f9aeb
Compare
Choose a tag to compare
Use system failure for node deletion or infra level issues (#59)

* Use system failure for node deletion or infra level issues

* using version

* removing PhaseSystemRetryableFailure

* code review feedback

* fixing a test

* idl version

* reverting tests

Fix AWS Batch CPU Scale

26 Feb 22:25
2c44d9b
Compare
Choose a tag to compare
Fix AWS Batch CPU Scale (#60)

* Fix AWS Batch CPU Scale

* PR Comments

Abort AWS Jobs on Finalize()

20 Feb 18:18
a931a40
Compare
Choose a tag to compare
Abort AWS Batch Job on Finalize (#58)

AWS Batch Plugin can determine that a task has failed if it detects it can no longer satisfy the failure ratio requirement. When this happens the array task node will be terminated then retried (if there are retry count available). This should trigger a cancellation to the AWS Job as well. 

It was originally written this way to match the old version of this code that, upon discussion, has been deemed a bad design. 

- [X] Need to ensure the cancellation API is idempotent (or add error check/ignore if it's not)
         Validated by running a Workflow and waiting for it to kick off an AWS Job then aborting it. That ended up calling Abort() then Finalize() and the second call was no-op

Fix for Batch timeouts

18 Feb 20:41
1a10fb5
Compare
Choose a tag to compare
v0.3.4

Fix empty timeout (#57)