RDSC-4633 How to perform HA failover

[ilianiliev-redis]

Ticket: https://redislabs.atlassian.net/browse/RDSC-4633

Document how to perform an HA failover test.

---

Note

Low Risk
Low risk documentation-only change; no product code or behavior is modified. Main risk is operational confusion if the iptables example is followed incorrectly during testing.

Overview
Adds a new doc page, ha-test.md, describing how to deliberately trigger and observe RDI HA failover by blocking traffic to the RDI database (via iptables) and watching operator logs, plus cleanup steps.

Updates the VM HA installation guide (install-vm.md) to point readers to the new failover testing instructions.

^{Written by Cursor Bugbot for commit 48e096a. This will update automatically on new commits. Configure here.}

[andy-stark-redis]

Just a few style suggestions, but otherwise LGTM.

[andy-stark-redis]

Does this mean that you expect dig to return more than one IP address for a hostname? (Presumably you only need to run the command once on the leader node.) If so, maybe say that explicitly in step 1, because it currently says "Identify the database IP", which makes it sound like there is only one address, but "IP" might potentially be plural here.

[ilianiliev-redis]

It is possible that one hostname points to multiple IPs, for example:

% dig +short example.com
104.18.26.120
104.18.27.120

The idea is that if we have a DNS Round-robin or other load balancing happening at DNS level, we need to block all IPs, to ensure that we can not connect to any of those.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

[ilianiliev-redis]

@andy-stark-redis I have also added RDI here to make it clear which database we are talking about.