Can we talk about hot failover to a second machine?

I've been required to add redundancy to one of my installations. Basically, an entire second Linux machine sitting there waiting for the first one to fail.My project has a bunch of data collecting node processes all talking to other networked devices and writing to a local mongo database. Since I have a could-based supervisor, that now gives me 3 nodes, so I can replicate mongo data fairly easily.So, what sort of tool or mechanism should I use to figure out when to tell the secondary server to start all the node processes, after noticing the primary has failed.I could write my own scripts, and use pm2 to start and stop things, but I was wondering if this problem has already been solved, by things like Heartbeat, or some clever npm package that I can't seem to successfully google for.

Submitted November 13, 2017 at 06:37PM by mainstreetmark