Question:
How to Reset inter-cluster replication - Hot initialization
Answer:
- Leave the Inter-Cluster replication Enabled and ensure it works correctly both ways. (you can make some dummy changes and validate that first)
- Export the HDAP store from the Central Journal Cluster into LDIF with “Export for replication” enabled
- Verify that the top few lines in the LDIF file should contain “vdssynccursor” as a date format and not a number
- Copy the LDIF file to the other Secondary Cluster
- Perform an initialization using the exported LDIF into the other Secondary Clusters
- Once all the data is synced, the secondary cluster HDAP store's “vdssynccursor” value will change from a date format to a number format with the latest data change.
NOTE: After initialization, converting "vdssynccursor" from a Date format to a Number Format can take a very long time, depending on the number of changes in the store. Until the conversion is finished, all the replication traffic will be queued/pending on this secondary cluster. If the format doesn't change, you need to disable and enable the inter-cluster replication on the secondary cluster to kick-start the sync.
Advantage
All changes that happened during the Export process will also be published in other Clusters.
Disadvantage
This is a One-way replication data fix from Primary to Secondary.
Comments
Does this apply to VDS 7.2, 7.3, 7.4 and 8.0?
It was added after v7.2.25+ and v7.3.2+ and v7.4.0+ and will surely be in 8.0 code base.
Please sign in to leave a comment.