Skip to content
This repository has been archived by the owner on Jan 24, 2024. It is now read-only.

Commit

Permalink
Fix spelling and grammar
Browse files Browse the repository at this point in the history
  • Loading branch information
dwoz committed Sep 6, 2023
1 parent 454c9d3 commit a00e53b
Showing 1 changed file with 10 additions and 10 deletions.
20 changes: 10 additions & 10 deletions 0000-master-cluster.md
Original file line number Diff line number Diff line change
Expand Up @@ -12,13 +12,13 @@ Add the ability to create a cluster of Masters that run behind a load balancer.
# Motivation
[motivation]: #motivation

The current [high availability features](https://docs.saltproject.io/en/latest/topics/highavailability/index.html) in the Salt ecosystem allow minions to have back up masters. There are two flavors of Multi Master which Minions can be configured.
The current [high availability features](https://docs.saltproject.io/en/latest/topics/highavailability/index.html) in the Salt ecosystem allow minions to have back up masters. There are two flavors of Multi Master which can be configured on a Minion.

Minions can connect to [multiple masters simultaneously](https://docs.saltproject.io/en/latest/topics/tutorials/multimaster.html).

<img src='/diagrams/000-multi-master.png' width='400px'>

Minions can also be configure to one master at a time [using fail over](https://docs.saltproject.io/en/latest/topics/tutorials/multimaster_pki.html#multiple-masters-for-a-minion).
Minions can also be configured to connect to one master at a time [using fail over](https://docs.saltproject.io/en/latest/topics/tutorials/multimaster_pki.html#multiple-masters-for-a-minion).

<img src='/diagrams/000-multi-master-failover.png' width='400px'>

Expand All @@ -35,18 +35,18 @@ It would be much more ideal if jobs could scale across multiple masters.
# Design
[design]: #detailed-design

In order to accomplish this we will need to change the way jobs execute.
Currently new jobs get send directly to the publish server from the request
In order to accomplish this, we will need to change the way jobs execute.
Currently new jobs get sent directly to the publish server from the request
server.

<img src='/diagrams/000-current-job-pub.png' width='400px'>

If we forward IPC Events between Masters we can get the return flow to be shared, as shown below.
If we forward IPC Events between Masters, we can get the return flow to be shared, as shown below:


<img src='/diagrams/000-cluster-job-pub.png' width='400px'>

To get job publishes to work we need to make sure publishes also travel over the IPC Event bus.
To get job publishes to work, we need to make sure publishes also travel over the IPC Event bus.


<img src='/diagrams/000-cluster-fwd.png' width='400px'>
Expand All @@ -57,7 +57,7 @@ Jobs can come and go through all the masters in our master pool. From a minion's
<img src='/diagrams/000-cluster-arch.png' width='400px'>

> [!IMPORTANT]
> THe current work for this SEP can be found [here](https://github.com/saltstack/salt/pull/64936)
> The current work for this SEP can be found [here](https://github.com/saltstack/salt/pull/64936)

## Alternatives
Expand All @@ -75,6 +75,6 @@ None as of this time.
# Drawbacks
[drawbacks]: #drawbacks

The biggest drawback is the fact that we'll need to maintain the three ways of
doing HA. This adds complexity however, if successfull. We can potential
depericate some of or all of the exiting HA functionality.
The biggest drawback is the fact that we will need to maintain three ways of
doing HA. This adds complexity however, if successfull. We can potentially
depericate some of, or all of, the exiting HA functionality.

0 comments on commit a00e53b

Please sign in to comment.