search

andrewrk / naught

Zero downtime deployment for your Node.js server using builtin cluster API
MIT License
793 stars 69 forks source link

Naught worker goes down suddenly #89

Closed shayaantx closed 8 years ago

shayaantx commented 8 years ago

Hello, cool lib, had a weird issue:

I've noticed in certain cases when I make a change that breaks my application (i.e., never sends process.send("online") in the new worker), the old worker dies for some reason, even though my understanding is that the old worker should stay up till the new worker is online.

The weird thing it doesn't happen on the first try, usually many tries later. My understanding is any number of failures before process.send("online") should NOT bring down the active/old worker.

Am I misunderstanding something here? Or is there more troubleshooting I can do?

I can understand if naught doesn't protect against a server killing itself, but this case seems strange

shayaantx commented 8 years ago

I'll debug this a bit more and see whats up