apache / apache/cloudstack

Enhance Disaster Recovery Scenario thru re-copy template/snapshot between zone

オープン
#8,672 コメント 1 件 リアクション 0 件 担当者 0 名 GitHub で見る
component:storage type:enhancement
主要言語
Java
スター
3.1k
フォーク
1.4k
平均マージ
6日 19時間
マージ済み PR(30日)
32

説明

##### ISSUE TYPE

* Enhancement Request

##### COMPONENT NAME

~~~
Snapshot
Tempalte
~~~

##### CLOUDSTACK VERSION

~~~
4.19
~~~

##### CONFIGURATION

##### OS / ENVIRONMENT

##### SUMMARY

While ACS 4.19 brought a new feature that enable copy disk snapshot to another zone, an idea came up to extend this feature to become a disaster recovery approach.

Assume administrator ensure all template and disk snapshots are already made the copy to the partner zone,
when the victim zone's primary/secondary storage is unusable or corrupted, we will have an opportunity to just copy the template/snapshot from partner zone back to victim zone after the storage were rebuilt. However since the current ACS did not design to handle such scenario, so the VM originally host on victim zone has to deploy as a new instance on partner zone and start a new lifecycle.

My preliminary idea is

Scenario 1 - When victim zone primary storage is dead and unrecoverable.
1. Rebuilt a new primary storage
2. ACS found the victim zone instance volume are unavailable
3. We revert the volume from the snapshot image reside on secondary storage (Full Clone).

Scenario 2 - When victim zone secondary storage is dead and unrecoverable.
1. Rebuild a new secondary storage
2. Implement replace copy mechanism for template/snapshot from partner zone to victim zone
3. When we try to revert a snapshot, ACS found the victim zone snapshot disk is lost in victim zone, ACS then copy the snapshot from partner zone to the fresh secondary storage and do the disk recovery (Full Clone).

Scenario 3 - When victim zone both primary secondary storage is dead and unrecoverable.
1. Rebuild both new primary and secondary storage
2. Implement replace copy mechanism for template/snapshot from partner zone to victim zone
3. ACS found the victim zone instance volume are unavailable
5. We revert the volume from the snapshot image
6. ACS found the victim zone snapshot disk is lost in victim zone, ACS then copy the snapshot from partner zone to the fresh secondary storage and do the disk recovery (Full Clone).

With such implementation, when ACS setup multiple zone with scheduled disk snapshot, it will facilitate recovery scenario itself without engaging third party backup solution.

##### STEPS TO REPRODUCE

~~~

~~~

##### EXPECTED RESULTS

~~~

~~~

##### ACTUAL RESULTS

~~~

~~~

コントリビューションガイド

コントリビューションガイドを開く

調査の方向性

Issue ではファイル、テスト、エントリポイントが特定されていません。まず、CloudStack の snapshot、template、storage、クロスゾーンコピーのワークフローをマッピングし、次に 3 つすべての disaster-recovery シナリオに必要なリカバリ動作とテストを定義してください。

索引モデルが issue の本文から書いたものです。

評価

技術スタック
java
領域
cloud, infrastructure
issue の種類
機能追加
難易度
5/5
見積もり時間
1週間以上
活発さ
停滞
明瞭さ
説明が足りない
初心者へのやさしさ
25/100

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。