[xCluster] Automatically Replicate the Indexes of a Table using xCluster Asynchronous Replication

[mrajcevic01]

Jira Link: DB-801

Description

While replicating tables using xCluster Asynchronous Replication, indexes of the replicated table do not get replicated over. They have to be replicated as part of the initial setup_universe_replication command or added after the fact using alter_universe_replication. This is an issue for scenarios where a single table may have a large number of indexes. We should automatically replicate indexes of the replicated table to the target cluster without requiring users to have to call out the indexes themselves.

[dotyjim-work]

[docdb] xCluster Roadmap · Issue #10319 · yugabyte/yugabyte-db

Has a section that talks about DDLs, but I don't see an item that talks specifically about flowing DDL changes, however platform has an issue that addresses this.

[Platform] xCluster replication management through Platform · Issue #8963 ¹

Allow safe DDL changes to be propagated automatically to target universes

But we do have some internal discussions taking place:

• XCluster + Metadata Alters - Google Docs
• Unidirectional 2dc + backfill proposals - Google Docs (relevant for tables with data already - see [xCluster] Integration with index backfill #7613)

Footnotes

1. PLAT-580 ↩