Gitlab Backups and Restoresπ
Gitlab Helm Chart Configurationπ
- Follow the
Backup and rename gitlab-rails-secret
task within the Production document. - Fill in our externalStorage values, specifically
addons.gitlab.objectStorage.iamProfile
or both.Values.addons.gitlab.objectStorage.accessKey
&.Values.addons.gitlab.objectStorage.accessSecret
along with.Values.addons.gitlab.objectStorage.bucketPrefix
or you can override in the name for your own bucket eg:addons: gitlab: values: global: appConfig: backups: bucket: "BUCKET_NAME"
- If you would like to perform manual backups, you will need to ensure the
tmp
location in the toolbox pod has a PVC attached:addons: gitlab: values: gitlab: toolbox: persistence: enabled: true size: 100Gi
Backing up Gitlabπ
Manual Stepsπ
To perform a manual complete backup of Gitlab, exec into your Gitlab Toolbox pod and run the following: 1. Find your Gitlab Toolbox pod.
kubectl get pods -l release=gitlab,app=toolbox -n gitlab
kubectl exec -it gitlab-toolbox-XXXXXXXXX-XXXXX -n gitlab -- /bin/sh
backup-utility --skip registry,lfs,artifacts,packages,uploads,pseudonymizer,terraformState,backups
You can read more on the upstream documentation: https://docs.gitlab.com/charts/backup-restore/backup.html#create-the-backup.
Automatic Cron-based Backupsπ
It is recommended to set up automatic backups via Gitlab toolbox’s cron settings:
addons:
gitlab:
values:
gitlab:
toolbox:
backups:
cron:
enabled: true
extraArgs: "--skip registry,lfs,artifacts,packages,uploads,pseudonymizer,terraformState,backups"
persistence:
enabled: true
size: '200Gi'
resources:
limits:
cpu: 800m
memory: "2Gi"
requests:
cpu: 800m
memory: "2Gi"
Restore Gitlabπ
- Ensure your gitlab-rails secret is present in gitops or in-cluster and it correctly matches the database to which the chart is pointed.
- If you need to replace or update your rails secret, once it is updated be sure to restart the following pods:
kubectl rollout -n gitlab restart deploy/gitlab-sidekiq-all-in-1-v2 kubectl rollout -n gitlab restart deploy/gitlab-webservice-default kubectl rollout -n gitlab restart deploy/gitlab-toolbox
- Exec into the toolbox pod and run the backup-utility command:
- find your Gitlab Toolbox pod.
kubectl get pods -l release=gitlab,app=toolbox -n gitlab kubectl exec -it gitlab-toolbox-XXXXXXXXX-XXXXX -n gitlab -- /bin/sh
- Find your most recent backup from cloud storage by finding the last line of your most recent backup job pod:
kubectl get po -l release=gitlab,job-name -n gitlab --sort-by=.metadata.creationTimestamp kubectl logs gitlab-toolbox-backup-XXXXXXXX-XXXXX -n gitlab
- Find your most recent backup via AWS CLI:
aws s3api list-objects --bucket gitlab-backups --query 'reverse(sort_by(Contents,&LastModified))[0].Key' --output # Save the output, it is in the format TIMESTAMP_VALUE.tar
- Execute the backup-utility command which will pull down the tarred data from your configured cloud storage and restore.
You can read more on the upstream documentation: https://docs.gitlab.com/charts/backup-restore/restore.html#restoring-the-backup-file.
# Using the filename backup-utility --restore -f s3://BUCKET_NAME/ARCHIVE_NAME.tar # Using the Timestamp backup-utility --restore -t TIMESTAMP_VALUE
Last update:
2024-07-18 by Michael Martin