Alex Sharp

Probably worth mentioning that this example no longer works: https://github.com/BinaryMuse/fluxxor/blob/master/examples/react-router/app/app.jsx.

Something else I've noticed is that the existing nginx workers take an unusually long amount of time to shut down when I run a `reload`. Like a good 30 seconds,...

@matsumotory Any ideas on this?

Yes, I can try. I can try wrapping the `mruby_ssl_handshake_handler_code` in a begin/rescue. I'm not sure how/why a reconnect would do anything there, but I'll try and report back.

@matsumotory Is there a reconnect interface or method on the mruby-redis client?

> Yes, it's strange behavior. Maybe OS or middleware layer settings disconnects the TCP connection I guess. @matsumotory We're using the same redis server (AWS elasticache) in our main app...

@matsumotory I tried using `#enable_keepalive` and it didn't help -- still getting the errors. I will try the begin rescue you linked to and report back.

I've put begin/rescue blocks in both the `mruby_ssl_handshake_handler_code` block and the `mruby_init_worker_code` block. Errors are only being generated from the handshake handler block. Here's my code: ```ruby begin ssl =...

@matsumotory Is it possible to add more context here to the error message? https://github.com/matsumotory/mruby-redis/blob/18b3d06cf238b26eb921e019fff1a9aa9ad22e95/src/mrb_redis.c#L1462

@yyamano I will try an upgrade and let you know. Thanks for checking in on this.