How to migrate Nexus from embedded database to external postgres DB on kubernetes platform – Sonatype Support

Nexus Repository Manager can be deployed in kubernetes platform, with embedded orient or h2 database, using Sonatype helm charts.

Note: We strongly recommend against running Nexus Repository 3 with embedded database within container orchestration environments such as Kubernetes. Doing so can lead to severe data corruption.

This article explains migrating Nexus pro instances to external DB

Prerequisites

Kubernetes 1.19+
kubectl CLI with admin access
Nexus Pro instance, with embedded DB.
External Postgresql DB for nexus instance
Adequate free space on persistent volume (minimum of Double the existing DB size or 10GB)
Familiarize high level steps from migration document

Identifying Nexus Deployment

Collect below details from existing nexus deployment

Namespace in which nexus is deployed (This example uses nx3)

Deployment name

$ kubectl get deploy -n nx3
NAME READY UP-TO-DATE AVAILABLE AGE
nex3-nexus-repository-manager 1/1 1 1 41m
sri-postgres 1/1 1 1 99d

Note down the Nexus image version and arguments used for the deployment

 spec:
   containers:
   - env:
     - name: INSTALL4J_ADD_VM_PARAMS
       value: |-
          -Xms2703M -Xmx2703M
          -XX:+UnlockExperimentalVMOptions
          -XX:+UseCGroupMemoryLimitForHeap
          -Djava.util.prefs.userRoot=/nexus-data/javaprefs
          -Dnexus.licenseFile=/etc/nexus-license/license.lic
     - name: NEXUS_SECURITY_RANDOMPASSWORD
       value: "true"
     image: sonatype/nexus3:3.42.0
     imagePullPolicy: IfNotPresent

Preparing for migration

Schedule adequate downtime for migration
Perform a full backup using normal backup procedures

Shut down Nexus Repository gracefully

$ kubectl -n nx3 scale --replicas=0 deploy/nex3-nexus-repository-manager
deployment.apps/nex3-nexus-repository-manager scaled

$ kubectl get deploy -n nx3
NAME READY UP-TO-DATE AVAILABLE AGE
nex3-nexus-repository-manager 0/0 0 0 79m
sri-postgres 1/1 1 1 99d

Create a patch file as below, replacing with local DB values and patch the deployment to stop nexus from starting

$ cat patch-file.yaml
spec:
  template:
    spec:
      containers:
      - name: nexus-repository-manager
        command: ["/bin/bash"]
        args: ["-c", "while true; do echo ; sleep 100;done"] 
        env:
          - name: INSTALL4J_ADD_VM_PARAMS
            value: "-Xms2703M -Xmx2703M -XX:+UnlockExperimentalVMOptions -XX:+UseCGroupMemoryLimitForHeap -Djava.util.prefs.userRoot=/nexus-data/javaprefs -Dnexus.licenseFile=/etc/nexus-license/license.lic -Dnexus.datastore.enabled=true -Dnexus.datastore.nexus.jdbcUrl=jdbc:postgresql://sri-postgres-svc:5432/nexusdb -Dnexus.datastore.nexus.username=nexus -Dnexus.datastore.nexus.password=nexus123"
        livenessProbe:
          failureThreshold: 600
        readinessProbe:
          failureThreshold: 600

$ kubectl -n nx3 patch deployment nex3-nexus-repository-manager --patch-file patch-file.yaml
deployment.apps/nex3-nexus-repository-manager patched

Alternately you edit the deployment directly as below (replacing with local values)
```
$ kubectl -n nx3 edit deployment nex3-nexus-repository-manager
deployment.apps/nex3-nexus-repository-manager edited
```
Also increase the failureThreshold of readiness probe and liveness probe, to avoid pod restarts during migration.
```
readinessProbe:
  failureThreshold: 600

livenessProbe:
failureThreshold: 600
```

Start the Deployment and login to nexus instance

$ kubectl -n nx3 scale --replicas=1 deploy/nex3-nexus-repository-manager
deployment.apps/nex3-nexus-repository-manager scaled

$ kubectl get pods -n nx3
NAME READY STATUS RESTARTS AGE
nex3-nexus-repository-manager-7c6f767dd5-sldq2 0/1 Running 0 43s
sri-postgres-5ff49c75f4-f85dc 1/1 Running 27 (21h ago) 99d

$ kubectl -n nx3 exec -it nex3-nexus-repository-manager-7c6f767dd5-sldq2 -- /bin/bash
bash-4.4$

Change to persistent volume and download the latest DB migrator

bash-4.4$ cd /nexus-data
bash-4.4$ pwd
/nexus-data

bash-4.4$ curl -s -L -O https://download.sonatype.com/nexus/nxrm3-migrator/nexus-db-migrator-3.42.0-02.jar
bash-4.4$ ls -l nexus-db-migrator-3.42.0-02.jar
-rw-r--r-- 1 nexus nexus 45324415 Nov 8 04:16 nexus-db-migrator-3.42.0-02.jar

Executing the migration

1. Start the migrator from nexus-data/db directory, replacing with local DB values


bash-4.4$ cd /nexus-data/db
bash-4.4$ java -Xmx4G -Xms4G -XX:MaxDirectMemorySize=4014M -jar ../nexus-db-migrator-3.42.0-02.jar --migration_type=postgres --db_url="jdbc:postgresql://sri-postgres-svc:5432/nexusdb?user=nexus&password=nexus123"
04:44:41 [main] WARN c.s.n.d.migrator.MigratorApplication - Please ensure any Nexus Repository instance has been gracefully shut down before proceeding.
04:44:41 [main] INFO c.s.n.d.migrator.MigratorApplication - Do you want to continue [y/n]?
y
04:44:42 [main] INFO c.s.n.d.migrator.MigratorApplication - --content_migration parameter is absent. Setting it to true.
04:44:42 [main] INFO c.s.n.d.migrator.MigratorApplication - Force parameter wasn't found. Setting it to false by default.
04:44:42 [main] INFO c.s.n.d.migrator.MigratorApplication - ------------------------------------------------------------
04:44:42 [main] INFO c.s.n.d.migrator.MigratorApplication - Java version: Red Hat, Inc. 1.8.0_345
04:44:42 [main] INFO c.s.n.d.migrator.MigratorApplication - JVM arguments: -Xmx4G -Xms4G -XX:MaxDirectMemorySize=4014M
04:44:42 [main] INFO c.s.n.d.migrator.MigratorApplication - Migrator arguments: --migration_type=postgres --db_url=jdbc:postgresql://sri-postgres-svc:5432/nexusdb?user=nexus&password=**** --content_migration=true --force=false
04:44:42 [main] INFO c.s.n.d.migrator.MigratorApplication - ------------------------------------------------------------
04:44:43 [main] INFO c.s.n.d.migrator.MigratorApplication - Starting MigratorApplication v3.42.0-02 using Java 1.8.0_345 on nex3-nexus-repository-manager-5bd8797685-rvtv5 with PID 37 (/nexus-data/nexus-db-migrator-3.42.0-02.jar started by nexus in /nexus-data/db)
04:44:43 [main] INFO c.s.n.d.migrator.MigratorApplication - No active profile set, falling back to 1 default profile: "default"
04:44:44 [main] INFO o.s.b.c.r.s.JobRepositoryFactoryBean - No database type set, using meta data indicating: POSTGRES
04:44:44 [main] INFO c.s.n.d.migrator.MigratorApplication - Started MigratorApplication in 1.668 seconds (JVM running for 3.576)
04:44:44 [main] INFO o.s.b.a.b.JobLauncherApplicationRunner - Running default command line with: [--migration_type=postgres, --db_url=jdbc:postgresql://sri-postgres-svc:5432/nexusdb?user=nexus&password=nexus123, --content_migration=true, --force=false]
04:44:44 [main] INFO c.s.n.d.m.validator.PathValidator - Validating Orient directory: /nexus-data/db
04:44:44 [main] INFO c.s.n.d.m.validator.PathValidator - Validating Orient directory: /nexus-data/db
04:44:44 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - Migration job started at Tue Nov 08 04:44:44 GMT 2022
04:44:44 [main] INFO o.s.batch.core.job.SimpleStepHandler - Executing step: [startHealthCheckDeciderStep]
04:44:44 [main] INFO o.s.batch.core.step.AbstractStep - Step: [startHealthCheckDeciderStep] executed in 24ms
04:44:44 [main] INFO o.s.batch.core.job.SimpleStepHandler - Executing step: [exportFilesStep]
04:44:44 [main] INFO c.s.n.d.m.export.JsonFilesExporter - Export may take several minutes depending on the amount of data
04:44:45 [main] INFO c.s.n.d.m.export.JsonFilesExporter - Exported config JSON file path /tmp/export_json-8995273030735844695/config
04:44:45 [main] INFO c.s.n.d.m.export.JsonFilesExporter - Exported security JSON file path /tmp/export_json-8995273030735844695/security
04:44:46 [main] INFO c.s.n.d.m.export.JsonFilesExporter - Exported component JSON file path /tmp/export_json-8995273030735844695/component
04:44:46 [main] INFO o.s.batch.core.step.AbstractStep - Step: [exportFilesStep] executed in 1s637ms
04:44:46 [main] INFO o.s.batch.core.job.SimpleStepHandler - Executing step: [readFileStep]
04:44:46 [main] INFO c.s.n.db.migrator.reader.FileReader - Current migration type is postgres.
04:44:46 [main] INFO c.s.n.db.migrator.reader.FileReader - Total number of config and security records to be processed = 36
04:44:46 [main] INFO o.s.batch.core.step.AbstractStep - Step: [readFileStep] executed in 98ms
04:44:46 [main] INFO o.s.batch.core.job.SimpleStepHandler - Executing step: [initDatabaseStep]
04:44:47 [main] INFO o.s.batch.core.step.AbstractStep - Step: [initDatabaseStep] executed in 498ms
04:44:47 [main] INFO o.s.batch.core.job.SimpleStepHandler - Executing step: [processConfigStep]
04:44:47 [main] INFO c.s.n.d.m.p.QuartzJobDetailProcessor - Filtered QuartzJobDetail record Cleanup service
04:44:47 [main] INFO c.s.n.d.m.p.QuartzJobDetailProcessor - Filtered QuartzJobDetail record Storage facet cleanup
04:44:47 [main] INFO c.s.n.d.m.p.QuartzTriggerProcessor - Filtered QuartzTrigger record Cleanup service
04:44:47 [main] INFO c.s.n.d.m.p.QuartzTriggerProcessor - Filtered QuartzTrigger record Storage facet cleanup
04:44:47 [main] INFO o.s.batch.core.step.AbstractStep - Step: [processConfigStep] executed in 100ms
04:44:47 [main] INFO o.s.batch.core.job.SimpleStepHandler - Executing step: [processComponentStep]
04:44:47 [main] INFO o.s.batch.core.step.AbstractStep - Step: [processComponentStep] executed in 116ms
04:44:47 [main] INFO o.s.batch.core.job.SimpleStepHandler - Executing step: [createRebuildRepositoryTasksStep]
04:44:47 [main] INFO o.s.batch.core.step.AbstractStep - Step: [createRebuildRepositoryTasksStep] executed in 11ms
04:44:47 [main] INFO o.s.batch.core.job.SimpleStepHandler - Executing step: [finalDatabaseStep]
04:44:47 [main] INFO o.s.batch.core.step.AbstractStep - Step: [finalDatabaseStep] executed in 26ms
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - Migration job finished at Tue Nov 08 04:44:47 GMT 2022
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - Migration job took 2 seconds to execute
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - 78 records were processed
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - 4 records were filtered
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - 0 records were skipped
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - 74 records were migrated
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - Created 'Rebuild repository browse' and 'Rebuild repository search' tasks. They will automatically one-time run after starting your Nexus Repository instance.
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - Migrated only early access formats: [APT, COCOAPODS, CONAN, CONDA, DOCKER, GITLFS, GO, HELM, MAVEN2, NPM, NUGET, P2, PYPI, R, RAW, RUBYGEMS, YUM]
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - Cleaning orient temporary files...
04:44:47 [main] INFO c.s.n.d.m.l.ProvidingJobInfoListener - Cleaning finished

Starting with external DB

Now that data is migrated to external database, stop the deployment and remove the command and args from the nexus deployment.

Stop Deployment

$ kubectl -n nx3 scale --replicas=0 deploy/nex3-nexus-repository-manager
deployment.apps/nex3-nexus-repository-manager scaled

Create a deployment patch file as below and execute kubectl patch with it to remove command and args

$ cat patch-remove.yaml
- op: remove
  path: "/spec/template/spec/containers/0/command"
- op: remove
  path: "/spec/template/spec/containers/0/args"

$ kubectl patch deploy nex3-nexus-repository-manager -n nx3 --patch-file patch-remove.yaml --type=json
deployment.apps/nex3-nexus-repository-manager patched

Start Nexus Deployment

$ kubectl -n nx3 scale --replicas=1 deploy/nex3-nexus-repository-manager
deployment.apps/nex3-nexus-repository-manager scaled
$ kubectl get pods -n nx3
NAME READY STATUS RESTARTS AGE
nex3-nexus-repository-manager-7fffb698d8-kgj7q 0/1 Running 1 (18s ago) 40s
sri-postgres-5ff49c75f4-f85dc 1/1 Running 27 (22h ago) 99d

3. Login to Nexus UI and confirm External DB is used

Prerequisites

Identifying Nexus Deployment

Preparing for migration

Executing the migration

Starting with external DB

Related articles