Upgraded to 7.9.1 this morning and had a bit of a shock when my data wasn't reloading. Finally RTFM and saw the note about dangling indexes no longer reloaded by default. Oh boy.
Anyway - when v8 rolls around - how are we to deal with clusters that get into a 'bad state' and have to be reloaded? I run everything on kubernetes and - while rarely - it does happen that I lose the master node and have to restart everything. This typically means reloading all the data and letting elastic sort it out.
What's the plan for such a scenario going forward?