release-24.0.0

This version provides a new CRD (DDC) to provide Apache Doris deployment capabilities based on compute-storage decoupled mode (an important new feature of Apache Doris 3.0).
In addition, it better supports the previous compute-storage coupled mode deployment (based on DCR).
Note that doris-operator compute-storage decoupled is only applicable to Apache Doris version >= 3.0.2.

New Features

Disaggregated Doris Cluster

• Support disaggregated doris fdb module access #186
• Support disaggregated doris metaService module deployment #194
• Support disaggregated doris fe module deployment #199
• Support disaggregated doris compute group module deployment #185 #189 #192 #197
• Support disaggregated doris for image entrypoint file #195 #196
• Support disaggregated doris fe scale down #225
• Support disaggregated doris compute group scale down and resource clear #238
• Support disaggregated doris fe scale down for sql interface #255
• Support disaggregated doris compute group scale down for sql interface #256

Doris Cluster

• Support Arrow Flight SQL Port in External Service #251

Improvements

Disaggregated Doris Cluster

• Support disaggregated doris compute group is created through be.conf and pvc. #198
• Readme file and DDC example yaml file #203 #210 #211 #214
• Readme example use latest version #220
• Add operator disaggregated switch to open DDC deploy #204
• Add SystemInitialization to pre-set system environment configuration #212
• Rename compute group to compute cluster following Doris's changes #215
• Rename compute cluster to compute group following Doris's changes #245
• Remove CRD for metaService replicas #227
• Reform ms(CRD) to cluster(CRD named DDC) #234
• Modify compute group identifier to UniqueId for new sql interface #239
• Support disaggregated fe sql entrypoint for docker image and add cluster id hash code #249
• Modify operator for the doris change from http interface to SQL interface #254
• Add pod liveness probe TimeoutSeconds for DDC #257
• Add logNotStore for fe to skip create log pvc #266
• Modify DDC example yaml file and disaggregated readme #268

Doris Cluster

• Add pod liveness probe TimeoutSeconds for DCR #257

Other

• Add unit test for common util code #226
• Add unit test for resource model and controller model #232
• Modify helm chart config to artifacthub.io repo #244
• Modify organize from Selectdb to Apache after the donation #247

Bug Fixes

Disaggregated Doris Cluster

• Fix pod affinity and statefulset pvc point bug #209
• Modify disaggregated be default storage path #243
• Fix pod restart when first deploying due to redundant env named ms_token #259
• Fix be configmap cannot use due to docker image entrypoint configmap path is wrong #261
• Fix electionNumber is nil to deploying cluster is wrong and service always updated #262 #266
• Fix the issue of reconcile triggering repeatedly #263 #265
• Fix the default path of be cache #266

Doris Cluster

• Fix upgrade operator restart doris pod #226
• Fix pod env to use deepcopy for splice dis-share array #236
• Fix mysql client failure when scaling down fe in different namespaces between operator and doris clusters #240
• Fix electionNumber is nil to deploying cluster err #260

Other

• Fix resource model code unit test for statefulset #252

Thanks contribute

• intelligentfu8
• catpineapple
• hechao-ustc

Downloads

• operator image
• doris image :
• • ms image
• • fe image
• • be image

1.6.1

1.6.1

this version updated CRD, Added a parameter about pod startup timeout. Note that this is a temporary version and helm has not adapted the changes. Avoid the phenomenon that upgrading the operator causes the doris cluster to restart.

Upgrade Tips

From version 1.4.1 to 1.6.0, it is recommended to upgrade directly to version 1.6.1 to avoid the problem of DorisCluster rolling restart caused by the upgraded operator.

Features

1. added a parameter about pod startup timeout. (pr: 190)
2. helm supported specify imagePullSecret for operator deployment. (pr 182)

Fixed

1. avoid the phenomenon that upgrading the operator causes the doris cluster to restart. (pr: 190)
2. fe scale down failed. (issue 205, pr 207)
3. upgrade golang version of coding for security.(issue 188, pr 193,191)

Downloads

• operator image
• doris image provided by selectdb:
• • fe image
• • be image
• • broker image

24.0.0-beta

Notice

This is an important milestone version for supporting deploy doris's separation of storage and compute mode.
You can reference the quickstart doc to deploy doris on separation of storage and compute mode.(Currently, doris does not support running on the arm64 platform)
Except supported deploy doris's separation of storage and compute mode, Doris-Operator fixed some bugs for deploying doris cluster.

Features

1. supported deploy separation of storage and compute mode.
   (pr 173,175,180,184,189,192,194,195,196,197,198,199,202,203,204,206,209,210,211,212,213,214,215,216,217,219,220)
2. helm supported specify imagePullSecret for operator deployment. (pr 182)

Fixed

1. fe scale down failed. (issue 205, pr 207)
2. upgrade golang version of coding for security.(issue 188, pr 193,191)
3. add a field for config pod start time, because pod may be need more time than default to restart.(pr 190)

Downloads

• operator image
• doris image provided by selectdb:
• • fe image
• • be image
• • ms image
• • broker image

timeout-fix

Notice

this version updated CRD, (base on 1.6.0)Added a parameter about pod startup timeout. Note that this is a temporary version(base on 1.6.0) and helm has not adapted the changes.

Upgrade Tips

From version 1.4.1 to 1.6.0 (current version is equivalent to 1.6.0), it is recommended to upgrade directly to version 1.6.1 to avoid the problem of DorisCluster rolling restart caused by the upgraded operator.

Features

1. Added a parameter about pod startup timeout. (pr: 190)

Downloads

• operator image
• doris image provided by selectdb:
• • fe image
• • be image
• • broker image

1.6.0

Notice

this version updated CRD, further refines the status of cluster phase, adds webhook, and provides automatic scaling down fe observer.

Upgrade Tips

From version 1.4.1 to 1.6.0, it is recommended to upgrade directly to version 1.6.1 to avoid the problem of DorisCluster rolling restart caused by the upgraded operator.

Features

1. add webhook ability. (pr: 168,170)
2. helm charts add webhook deployment. (pr: 171)
3. support msyql client for operator connect doris. (pr: 172)
4. support graceful scale down of fe observer. (pr: 174)

Fixed

1. restrict adding certificate in webhook enable. (pr: 169)
2. no longer print reconciler watch pod logs. (pr: 181)

Downloads

• operator image
• doris image provided by selectdb:
• • fe image
• • be image
• • broker image

1.5.2

Upgrade Tips

From version 1.4.1 to 1.6.0, it is recommended to upgrade directly to version 1.6.1 to avoid the problem of DorisCluster rolling restart caused by the upgraded operator.

Fixed

1. remove duplicate code for envs.162
2. fix spelling error.163
3. add empty value for helm configmap.164
4. delete all-fields doc.166

1.5.1

Notice

this version implemented some enhancements including BE exit gracefully, more granular pod/container security contexts, bug fixes for Service, session affinity, annotations to PVC, operator config imagePullPolicy, etc.

Upgrade Tips

From version 1.4.1 to 1.6.0, it is recommended to upgrade directly to version 1.6.1 to avoid the problem of DorisCluster rolling restart caused by the upgraded operator.

Features

1. add support for defining more granular pod/container security contexts.(pr: 141)
2. add option to specify which namespace to reconcile objects.(pr: 142)
3. add set terminationGracePeriodSeconds and add --grace in be_prestop.sh. (issues: 145, pr: 157)

Fixed

1. fix Service bug in lower versions. (issues: [149](#149, 143, pr: 160)
2. add annotations to PVC. (issues: 150, pr: 160)
3. add sessionaffinity for service. (issues: 155, pr: 160)
4. fix config operator imagePullPolicy.(pr: 159)

Removed

1. delete docker dir and move to Apache Doris Repo. (pr: 158, 33329)
   The relevant scripts have been merged into Apache Doris Repo

Downloads

• operator image
• doris image provided by selectdb:
• • fe image
• • be image
• • broker image

1.5.0

Notice

this version updated CRD, supporting mount configmaps for configurations, dangerous behavior restrictions on expansion and contraction operations, helm enhancements, etc.

Upgrade Tips

From version 1.4.1 to 1.6.0, it is recommended to upgrade directly to version 1.6.1 to avoid the problem of DorisCluster rolling restart caused by the upgraded operator.

Features

1. add recovery ability. (issues: 116, pr: 123)
2. scale down followers are not allowed. (issues: 117, pr: 128)
3. supports multiple configmaps for mounting custom directories. (issues: 118, pr: 135)
4. helm supports multiple configmaps for mounting custom directories.(pr: 140)
5. helm supports secret to config username and password.(pr: 140)
6. helm create namespace policy adjustment for install and upgrade doris-operator.(pr: 147)

Fixed

1. some apiserver will check the resourceVersion when update. (issues: 119, pr: 121)
2. add annotations for pod.(issues: 137, pr: 139)

Downloads

• operator image
• doris image provided by selectdb:
• • fe image
• • be image
• • broker image

1.4.1

Notice

this version have update startupProbe in order to consistent with livenessProbe.

Upgrade Tips

From version 1.4.1 to 1.6.0, it is recommended to upgrade directly to version 1.6.1 to avoid the problem of DorisCluster rolling restart caused by the upgraded operator

Features

1. use column name to find master to avoid output sequence changed. 111
2. add basic-auth type secret to doriscluster to config mangement username and password. 115

Fixed

1. fix helm workflows for update latest files in repository. 100
2. example name misspelling.104
3. compatible with chages in output sequence about IsMaster.105
4. add default edit_log_port for fe_entrypoint to use.106
5. check the operator deployed namespace exists or not to avoid create cr failed.109
6. when resourceVersion not set update service failed.119

Downloads

• operator image
• doris image provided by selectdb:
• • fe image
• • be image
• • broker image

1.4.0

Notice

this version have already supports deployment by Helm.

Features

1. add helm.62
2. add helm action for auto release.88
3. add helm namespace automatically.89

Fix

1. when update service type not effect.98
2. helm modify. 77
3. helm doirscluster abstract. 81
4. helm doriscluster split configmap. 90
5. helm readme and comments modify. 91
6. helm pvc storageclass modify. 92
7. helm fix helm clusterrole . 94

Downloads

• operator image
• doris image provided by selectdb:
• • fe image
• • be image
• • broker image
• Helm Charts provided by selectdb:
• • Doris-Operator Helm
• • DorisCluster Helm
• • Helm Repo