Hacktoberfest 2026: những issue maintainer đã đánh dấu cho tháng Mười, đang mở và phù hợp người mới. Xem issue Hacktoberfest

Record the object store and server name of a backup in Backup.status.pluginMetadata

Đang mở
  • Có gì không ổn?
#1,140 0 bình luận 0 reaction 0 người được giao Xem trên GitHub

Maintainer thường phản hồi trong vòng 1 ngày

@BoxBoxJason đang làm issue này rồi.

Từ ngày 4/10/2026.

  • #1141 của @BoxBoxJason — đang mở

Đánh giá

Độ khó
4/5
Thời gian dự kiến
3-5 ngày
Mức phù hợp với người mới
35/100
Loại issue
Tính năng
Độ rõ ràng
Đặc tả rõ ràng
Mức độ hoạt động
Đình trệ
Công nghệ
go, postgresql
Lĩnh vực
databases

Hướng nghiên cứu

Start in internal/cnpgi/instance/types.go at newBackupResultMetadata, then trace how the plugin resolves the ObjectStore and serverName for a backup. Check the existing backup/plugin tests and add coverage for the proposed metadata keys, including the optional endpointURL. Done means Backup.status.pluginMetadata records the backup location and resolved server name without changing existing metadata.

Do mô hình lập chỉ mục viết ra từ nội dung của issue.

Mô tả

Description

A Backup taken through the plugin doesn't record which ObjectStore and which serverName the backup was written to. When creating a recovery cluster, we have to find that out by other means and write it by hand in the externalClusters entry.
I'd like the plugin to add this information to the backup result metadata, so that it ends up in Backup.status.pluginMetadata.

This is related to #510 and cloudnative-pg/cloudnative-pg#7989 (recovery from a plugin Backup object), which I believe need this as a first step.

Our use case

Our clusters archive under a timestamped serverName (for example mydb-20261003T120000), set through the serverName plugin parameter, and a new one is generated for each new generation of a cluster.
The archive path has to be empty when a cluster starts archiving, and a fresh, dated serverName gives us that without deleting older archives.
Over time, the bucket holds one folder per generation.

To create a recovery cluster from a given backup, we need to know which of these folders holds it. The plugin's recovery default for serverName is the name of the new cluster (internal/cnpgi/operator/config/config.go), which never matches a timestamp, so we have to find the value by listing the bucket or
digging through old manifests.
The Backup object is the natural place to keep the link between a backup and its location, and it's where the in-tree integration kept it: Backup.status.serverName, destinationPath and endpointURL were all set. Since the in-tree integration is deprecated since 1.26, anyone migrating to the plugin loses that.

Environment
  • CloudNativePG operator: main (the behaviour is the same on current releases)
  • plugin-barman-cloud: v0.15.1
What a plugin backup records today
apiVersion: postgresql.cnpg.io/v1
kind: Backup
spec:
  method: plugin
  pluginConfiguration:
    name: barman-cloud.cloudnative-pg.io
status:
  backupId: ...
  pluginMetadata:
    clusterUID: ...
    displayName: ...
    name: ...
    pluginName: ...
    timeline: "1"
    version: ...

Nothing in it says where the backup went. newBackupResultMetadata (internal/cnpgi/instance/types.go) is where these keys are built.

Proposal

Add the configuration the backup was taken with to the same map:

Key Value
barmanObjectName name of the ObjectStore used for the backup
serverName resolved server name (the serverName parameter, or the cluster name when unset)
destinationPath ObjectStore.spec.configuration.destinationPath at backup time
endpointURL ObjectStore.spec.configuration.endpointURL at backup time, when set

This only adds keys and doesn't change any existing behaviour. destinationPath and endpointURL aren't strictly needed to locate the data, since they can be read from the ObjectStore, but they let a later restore detect that the ObjectStore has been changed to point somewhere else since the backup was taken.

Shipping this early matters even if nothing consumes it yet: backups taken before it lands will never carry the information.

Alternatives considered
  • Dedicated fields in Backup.status (serverName, destinationPath): they would need new fields in the CNPG-I BackupResult message for concepts that are specific to Barman, so pluginMetadata looks like the right place.
  • Keep the status quo and document that serverName must always be set in the recovery source.
    This is what we do today, but with timestamped names the value has to be found outside the cluster.
  • Change the default recovery serverName: no default can guess a timestamp, and changing it would break users who rely on the current one.

I used an AI assistant (Claude) to help draft this issue and to check the references against the plugin and operator sources, as the AI policy asks for disclosure.

Ngôn ngữ chính
Go
Star
196
Fork
76
Merge trung bình
4 ngày 13 giờ
Pull request đã merge (30 ngày)
19

Chuẩn bị môi trường

Bắt đầu từ đâu

  1. Đọc hết issue, rồi đọc hướng dẫn đóng góp của dự án.
  2. Bình luận trên issue rằng bạn sẽ nhận — tránh hai người làm cùng một việc.
  3. Fork repository và làm thay đổi trên một nhánh.
  4. Mở pull request có tham chiếu số hiệu của issue.

Issue khác của cloudnative-pg/plugin-barman-cloud

Tất cả issue của cloudnative-pg/plugin-barman-cloud

Issue tương tự

Thêm issue về Go

Nhận issue mới trong hộp thư của bạn

Bản tóm tắt ngắn những issue GitHub phù hợp với người mới.