I was performing a manual rolling update of my 3 node cluster when suddenly
one of my nodes went into "standalone" mode. I had never seen this mode
The short answer is that exhibitor was running on a node that I did a
"zkServer.sh restart" on...and it decided to "help" me by rewriting my
zoo.cfg. The problem is that it rewrote it as a single node cluster.
Is there any way to prevent this not-so-helpful behavior?
>From the netflix exhibitor page:
"*Each Exhibitor instance monitors the ZooKeeper server running on the same
server. If ZooKeeper is not running (due to crash, etc.), Exhibitor will
rewrite the zoo.cfg file and restart it*."