Rails 500 error without writing production log

Question

I installed a new rails application on the same server as another application. The original works fine, but the new application gives me the infamous "We're sorry, but something went wrong," but there is no entry in the journal. Recent log entries are from my migration using rake.

I found that if you run run cap: it's cold, it will work, but then on the next yupdate it will work again without a sock file. Then cover deployment: cold is required for each update.

edit: If I run /etc/init.d/unicorn_taxidata restart, I get "can not restart", then if I run it again, it will start normally.

The application is in development. The environment is NGINX, unicorn, postgresql, rails 4.0.0.0, ruby ​​2.0.0p195.

I have this error in the nginx error log:

[crit] 889#0: *65 connect() to unix:/tmp/unicorn.myapp.sock failed (2: No such file or directory) while connecting to upstream, client: 1.123.13.26, server: myapp.com.au, request: "GET /login HTTP/1.1", upstream: "http://unix:/tmp/unicorn.myapp.sock:/login"