1 個回答
- 最新
- 最多得票
- 最多評論
1
Handling Replication Slots The requirement to remove logical replication slots before upgrading poses a challenge for continuous replication. However, you’re not necessarily required to move the whole dataset from the beginning after the upgrade. Consider the following:
- Stop Replication: Temporarily stop the replication process in DMS and ensure all changes are applied to the target database to prevent data loss.
- Remove Replication Slots: As required by RDS, remove the logical replication slots to proceed with the upgrade.
- Upgrade: Perform the major version upgrade on your RDS instance.
After the Upgrade
- Re-create Replication Slots: After the upgrade, re-create the logical replication slots. This step is crucial for resuming logical replication.
- Resume or Restart Replication:
- Resume: In some cases, you might be able to resume replication by pointing to the new replication slot, assuming you can manage the sequence of changes accurately.
- Full Load + CDC: More commonly, you’ll initiate a new replication task in DMS, starting with a full load and transitioning into CDC. This doesn't necessarily mean moving the entire dataset from scratch if you manage to synchronize the changes made during the upgrade process.
相關內容
- 已提問 10 個月前
AWS 官方已更新 2 年前- AWS 官方已更新 7 個月前
- AWS 官方已更新 2 年前
Yeah. Thanks. It kind of works to drop replication slot and then re-create it after major version upgrade. But it requires that application is prevented accessing database between dropping and re-creating the slot. Otherwise data is lost in replication. Also it seems that blue-green deployment won't work with DMS so that can't be used either.