Remove support for delaying state recovery pending master nodes

[DaveCTurner]

It is useful to be able to delay state recovery until enough data nodes have joined the cluster, since this gives the shard allocator a decent opportunity to re-use as much existing data as possible. However we also have the option to delay state recovery until a certain number of master-eligible nodes have joined, and this seems unnecessary: we require a majority of master-eligible nodes for state recovery, but there is no advantage in waiting for more.

We should deprecate the following settings in 7.x and remove them in 8.0:

• gateway.expected_nodes
• gateway.expected_master_nodes
• gateway.recover_after_nodes
• gateway.recover_after_master_nodes

[elasticmachine]

Pinging @elastic/es-distributed (:Distributed/Cluster Coordination)

[DaveCTurner]

We discussed this today and agreed that it will help to simplify this area, but this isn't on the roadmap yet so I'm marking it as help wanted to invite contributions from the community.

[aldenpeffer]

@DaveCTurner I'm going to take a crack at this.

[DaveCTurner]

Thanks @aldenpeffer, go ahead. Note that this'll take two PRs in sequence: the first to deprecate these settings and the second to remove them, and both should be opened against the master branch. We will take care of backporting the first.

[vishalthakur1972]

you like a genius to me. can you convert this script to an amp script? plzzzz

<script type="text/javascript"> setTimeout(function(){ setInterval(function(){ gtag('event', 'Open', {'event_category': 'TimeOnPage'}); }, 10000); setInterval(function(){ if(!document.hidden) gtag('event', 'Visible', {'event_category': 'TimeOnPage'}); }, 10000); }, 60000); </script>