From 34a5fef2f7be69552257eaed238323f9ccb3beb6 Mon Sep 17 00:00:00 2001 From: Peter Bacsko Date: Mon, 29 Jul 2024 14:46:45 +0200 Subject: [PATCH 1/2] [YUNIKORN-2708] Release notes for 1.5.2 --- src/pages/release-announce/1.5.2.md | 35 +++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) create mode 100644 src/pages/release-announce/1.5.2.md diff --git a/src/pages/release-announce/1.5.2.md b/src/pages/release-announce/1.5.2.md new file mode 100644 index 00000000000..a8d36112f81 --- /dev/null +++ b/src/pages/release-announce/1.5.2.md @@ -0,0 +1,35 @@ +--- +id: rn-1.5.2 +title: Release Announcement v1.5.2 +--- + + + +# Release Announcement v1.5.2 +We are pleased to announce that the Apache YuniKorn community has voted to release 1.5.2. +Apache YuniKorn is a standalone resource scheduler, designed for managing and scheduling Batch and Data workloads on container +orchestration frameworks like Kubernetes for on-prem and on-cloud use cases. + +## Overview +The Apache YuniKorn community has resolved 17 [JIRAs](https://issues.apache.org/jira/issues/?filter=12353487) in this release. + +Release manager: Peter Bacsko + +Release date: 2024-07-26 From aa393808100a23f1a14a67583cd8c003e3a6292a Mon Sep 17 00:00:00 2001 From: Peter Bacsko Date: Tue, 30 Jul 2024 09:37:51 +0200 Subject: [PATCH 2/2] Review --- src/pages/release-announce/1.5.2.md | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/src/pages/release-announce/1.5.2.md b/src/pages/release-announce/1.5.2.md index a8d36112f81..a887dcbc661 100644 --- a/src/pages/release-announce/1.5.2.md +++ b/src/pages/release-announce/1.5.2.md @@ -33,3 +33,13 @@ The Apache YuniKorn community has resolved 17 [JIRAs](https://issues.apache.org/ Release manager: Peter Bacsko Release date: 2024-07-26 + +## Highlights +### Locking fix to avoid deadlock on node addition +Yunikorn could enter a deadlock when a node is added during scheduling. This is now fixed, see [YUNIKORN-2629](https://issues.apache.org/jira/browse/YUNIKORN-2629) for details. + +### CSINodes informer +Node volume limits and other that scheduling actions that relied on the CSINode objects failed as the CSINode informer was not started. This was fixed in [YUNIKORN-2621](https://issues.apache.org/jira/browse/YUNIKORN-2621) and the CSINode informer now gets started by default. + +### Skip setting the default queue name +The default queue name no longer gets set on the pod to allow placement rules to be executed correctly in specific configurations. Placement behaviour has not changed except for the specific configurations. Details are in [YUNIKORN-2703](https://issues.apache.org/jira/browse/YUNIKORN-2703) and [YUNIKORN-2711](https://issues.apache.org/jira/browse/YUNIKORN-2711).