Update sections that talk about "adding more master nodes"

[ozgune]

A prospective user asked about "adding more master nodes" today.

We have two sections in our documentation talk about adding master nodes. We should clarify our language in these sections to make sure that this isn't production-ready. We could either cover this type of advice in a separate section (more experimental) or hint at the upcoming MX solution.

We currently talk about adding master nodes in the following sections:

• Cluster Management > Scaling out your cluster > Adding a master (second paragraph)
• Scaling out data ingestion > Hash distributed tables > Masterless Citus (the example in this section is a bit out of sync with the previous sections)

[ozgune]

Email thread for reference:

mtuncer@ wrote:

Citus does not yet support inserting from multiple nodes. There is an ongoing project to add that support in near future.

We have a public user group at https://groups.google.com/forum/?utm_medium=email&utm_source=footer#!forum/citus-users. It contains discussions to other questions our users are asking.

We invite you to join that group, we hope you will find it useful.

The user replied:

I just read from the documents as below, thank you. I got it..

Adding a master

The Citus master only stores metadata about the table shards and does not store any data. This means that all the computation is pushed down to the workers and the master does only final aggregations on the result of the workers. Therefore, it is not very likely that the master becomes a bottleneck for read performance. Also, it is easy to boost up the master by shifting to a more powerful machine.

However, in some write heavy use cases where the master becomes a performance bottleneck, users can add another master. As the metadata tables are small (typically a few MBs in size), it is possible to copy over the metadata onto another node and sync it regularly. Once this is done, users can send their queries to any master and scale out performance. If your setup requires you to use multiple masters, please contact us at [email protected].

[skygiser]

This feature is very important in my option...

[ozgune]

We agree and we have a project in the works.

If you'd like to see a preview, we'd be happy to schedule a demo. I'm also cc'ing @marcocitus in this issue to keep him in the loop.

[begriffs]

http://docs.citusdata.com/en/v8.1/admin_guide/cluster_management.html#adding-a-coordinator

The second paragraph, starting with "However, in some write heavy use cases" could be improved to suggest using a Cloud follower for scaling reads, and MX scaling ingestion. Also include a link to the MX section for other use cases.