Looking at the replicated leveldb setup, it seems like it wouldn't handle partial failures.
For example, disk full, errors when trying to send messages to the queue, etc. So I assume proper management of a replicated leveldb setup would be to remotely kill the box if your monitoring detects failure. Shut down the ethernet port, halt the box, etc. This way the failover protocol kicks in…. Is this correct? -- Founder/CEO Spinn3r.com Location: *San Francisco, CA* blog: http://burtonator.wordpress.com … or check out my Google+ profile <https://plus.google.com/102718274791889610666/posts> <http://spinn3r.com>
