From 567c6bed9dee71c13163e83b3f92a0d728012e43 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Moritz=20Wanzenb=C3=B6ck?= Date: Mon, 20 Nov 2023 09:24:02 +0100 Subject: [PATCH] docs: highlight v2 changes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Moritz Wanzenböck --- docs/how-to/upgrade/index.md | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) diff --git a/docs/how-to/upgrade/index.md b/docs/how-to/upgrade/index.md index c9513b38..4c40aab8 100644 --- a/docs/how-to/upgrade/index.md +++ b/docs/how-to/upgrade/index.md @@ -30,3 +30,31 @@ This guide assumes: [Step 2:]: ./2-collect-information.md [Step 3:]: ./3-remove-operator-v1.md [Step 4:]: ./4-install-operator-v2.md + +## Key Changes Between Operator V1 and V2 + +### Administrative Changes + +* The resources `LinstorController`, `LinstorSatelliteSet` and `LinstorCSIDriver` have been replaced by + [`LinstorCluster`](../../reference/linstorcluster.md) and + [`LinstorSatelliteConfgiuration`](../../reference/linstorsatelliteconfiguration.md). +* The default deployment runs the LINSTOR Satellite in the [container network](../drbd-host-networking.md). + The migration script will propose changing to the host network. + +### Operational Changes + +* In Operator v1, all labels on the Kubernetes node resource where replicated on the satellite, making them usable in the + `replicasOnSame` and `replicasOnDifferent` parameters on the storage class. In Operator v2, only the following + labels are automatically synchronized. + * `kubernetes.io/hostname` + * `topology.kubernetes.io/region` + * `topology.kubernetes.io/zone` + Use [`LinstorSatelliteConfiguration.spec.properties`](../../reference/linstorsatelliteconfiguration.md#specproperties) + to synchronize additional labels. +* The following settings are applied by Operator v2 cluster-wide: + * DrbdOptions/Net/rr-conflict: retry-connect + * DrbdOptions/Resource/on-suspended-primary-outdated: force-secondary + * DrbdOptions/Resource/on-no-data-accessible: suspend-io + * DrbdOptions/auto-quorum: suspend-io +* Operator v2 also includes a [High-Availability Controller](https://github.com/piraeusdatastore/piraeus-ha-controller) + deployment to prevent stuck nodes caused by suspended DRBD devices.