How to resolve Rack Lint errors?

Question

I have noticed that certain requests, particularly from Uptime Robot to test page uptime have recently started generated the following Rack Lint error on my Rails 3.1.x app:

21:47:05 web.1     | Started HEAD "/" for 74.86.158.106 at 2012-03-22 21:47:05 -0400
21:47:05 web.1     |   Processing by HomeController#index as */*
...
21:47:05 web.1     | /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/activesupport-3.1.4/lib/active_support/core_ext/string/output_safety.rb:23: warning: regexp match /.../n against to UTF-8 string
...
21:47:05 web.1     | Completed 200 OK in 81ms (Views: 60.3ms | ActiveRecord: 0.0ms)
21:47:05 web.1     | 74.86.158.106 - - [22/Mar/2012 21:47:05] "GET / HTTP/1.1" 200 31842 0.1993
21:47:05 web.1     | E, [2012-03-22T21:47:05.546848 #18938] ERROR -- : app error: Content-Length header was 31842, but should be 0 (Rack::Lint::LintError)
21:47:05 web.1     | E, [2012-03-22T21:47:05.546978 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/rack-1.3.6/lib/rack/lint.rb:19:in `assert'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547079 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/rack-1.3.6/lib/rack/lint.rb:501:in `verify_content_length'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547178 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/rack-1.3.6/lib/rack/lint.rb:525:in `each'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547277 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/rack-1.3.6/lib/rack/body_proxy.rb:23:in `method_missing'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547375 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/unicorn-4.2.0/lib/unicorn/http_response.rb:41:in `http_response_write'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547476 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/unicorn-4.2.0/lib/unicorn/http_server.rb:538:in `process_client'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547574 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/unicorn-4.2.0/lib/unicorn/http_server.rb:604:in `worker_loop'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547672 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/unicorn-4.2.0/lib/unicorn/http_server.rb:487:in `spawn_missing_workers'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547770 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/unicorn-4.2.0/lib/unicorn/http_server.rb:137:in `start'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547867 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/gems/unicorn-4.2.0/bin/unicorn:121:in `<top (required)>'
21:47:05 web.1     | E, [2012-03-22T21:47:05.547966 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/bin/unicorn:19:in `load'
21:47:05 web.1     | E, [2012-03-22T21:47:05.548063 #18938] ERROR -- : /Users/username/.rvm/gems/ruby-1.9.3-p125@project/bin/unicorn:19:in `<main>'

Any ideas as to what may be going on here? I'm really not finding a solution right now that specifically addresses what to do in this scenario as far as poking and hunting down what really may be going on.

Answer 1

I know it's an old question, but I have the exact same issue. Requests from UptimeRobot and New Relic (and only from IPs from those two) give these log messages.

Both UptimeRobot and NewRelic reports 'OK'.

It seems that it is not the request that Rack::Lint is complaining about, but the response. Look at the stacktrace. It comes through http_response_write of unicorn (I too use unicorn).

I will investigate more and add any findings here.

UPDATE:

I am able to reproduce the issue with

$ curl https://example.com -X HEAD -i
HTTP/1.1 200 OK
Server: nginx/1.6.2 (Ubuntu)
Date: Wed, 18 Nov 2015 09:17:11 GMT
Content-Type: text/html; charset=utf-8
Content-Length: 18422
Connection: keep-alive
Status: 200 OK
...

curl: (18) transfer closed with 18422 bytes remaining to read

So it seems that my old Rails is not supporting HEAD requests properly.

UPDATE: I am not able to reproduce on my vagrant box or on the staging server. Only difference from production I can think of is SSL.