Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Evolution of Code Deploys at Reddit

redditblog.com

Post reply on HN
Print view View original

31–32 of 32 posts

Re: The Evolution of Code Deploys at Reddit

#31

Post by spladug »

syncopate wrote: post #20
spladug wrote: post #3

Earlier quoted context omitted.

Thanks! Do feel free to ask if you think of something :)

When you deploy: Will you lose requests that are currently being processed while you restart the service? Also, will a server still receive requests while the new code is being started?

Generally: no. The worker processes will finish up their current request before shutting down and being reaped. At the whole server level, Einhorn will indeed make sure that requests are still being served as the workers get shuffled out.

Re: The Evolution of Code Deploys at Reddit

#32

Post by gusfoo »

spladug wrote: post #30
pacaro wrote: post #22

Earlier quoted context omitted.

How do you handle one way gates? Clearly at each deploy there are two different versions running concurrently, and as you make changes you do so knowing that, but as the system evolves their are points in (code) time that you can't go back to. Is this not a concern because you would never roll back that far?

Yeah, rollbacks are more of adding a revert to the top of the pile so we just make sure we roll back things that can be rolled back. This is important to think about when planning deploys.

> Yeah, rollbacks are more of adding a revert to the top of the pile

Neat. It reminds me of the method used in gaming nowadays, which is to just write a "savepoint" marker in to the message stream instead of pausing the entire game to save state.

Post reply on HN