I've written a custom notification to Airbrake from a Ruby EM server we run along with our Rails app. In the case of an exception, I notify Airbrake as per the API docs
Airbrake.notify(
:error_class => "Ruby EM Server Error:",
:error_message => "EM Server Crashed with error: #{$!}",
:backtrace => $#,
) if ENV['RAILS_ENV'] == 'production'
However, when the error comes through as "Unknown: EM Server Crashed with error:"
I know what the RAILS_ENV is when I send off the notification so I was wondering if there was any way I can pass this info on to Airbrake without writing XML and posting it.
I'd like to see the error to come through as "Production: EM Server Crashed with error:"
Try this:
Airbrake.notify(
:error_class => "Ruby EM Server Error:",
:error_message => "EM Server Crashed with error: #{$!}",
:backtrace => $#,
:environment_name => ENV['RAILS_ENV']
) if ENV['RAILS_ENV'] == 'production'
The environment_name option is in the gem's documentation.
Related
I want to enable action caching in my rails app on heroku.
In development.rb I set:
config.action_controller.perform_caching = true
and see in logs
Started GET "..." for 127.0.0.1 at 2013-05-17 14:03:25 +0400
...
Write fragment ...
OR
Read fragment ... (0.2ms)
->
To move to production I installed memcache add-on via $heroku addons:add memcache , installed new gem in Gemfile: gem 'dalli' and changed settings in production.rb:
config.action_controller.perform_caching = true
config.cache_store = :dalli_store #, ENV['MEMCACHE_SERVERS'], { :namespace => 'myapp', :expires_in => 1.day, :compress => true }
I have also tried to enable those two commented parameters, but anyway I don't see Read/Write fragment ... pieces in logs, I see that app gets authenticated, but cache is always missing
Started GET "..." for 195.178.108.38 at 2013-05-17 09:54:19 +0000
Dalli/SASL authenticating as myapp%40heroku.com
Dalli/SASL: Authenticated
cache: [GET ...] miss
Running $heroku run console I check that the cache is loading:
irb(main):001:0> Rails.cache.read('color')
Dalli/SASL authenticating as myapp%40heroku.com
Dalli/SASL: Authenticated
=> nil
irb(main):002:0> Rails.cache.write('color', 'red')
=> true
irb(main):003:0> Rails.cache.read('color')
=> "red"
Why action caching does not work?
Can you try using memcachier instead?
remove memcahe add-on
add memcachier add-on
add "memcachier" gem just above "dalli" in your gem file
it should "just work"
See here in the DevCenter: Memcachier
I am attempting to use redis store as my session store on heroku. It works fine in development but I am unable to get a redis connection on heroku. It is attempting to connect to a 127.0.0.1 instead of the correct redis server.
Error:
ActionView::Template::Error (Error connecting to Redis on 127.0.0.1:6379 (ECONNREFUSED)):
I have set the heroku redis config to the correct server (not really using for the session store but it is set)
REDISTOGO_URL: redis://redistogo:#################################carp.redistogo.com:9274/
session_store.rb
GrnAuth::Application.config.session_store :redis_store, :server => APP_CONFIG['redis_server'], key: '_grn_session'
environment.rb
# Load the rails application
require File.expand_path('../application', __FILE__)
require 'yaml'
APP_CONFIG = YAML.load_file("#{Rails.root}/config/config.yml")[Rails.env]
# Initialize the rails application
GrnAuth::Application.initialize!
config.yml
development:
redis_server: redis://localhost:6379/
test:
redis_server: redis://localhost:6379/
production:
redis_server: redis://redistogo:#################################carp.redistogo.com:9274/
When I console in to heroku I can check APP_CONFIG['redis_server'] and it is set to the redis server.
I have also set up a redis connection just to use that works.
redis.rb
uri = URI.parse(ENV["REDISTOGO_URL"] || "redis://localhost:6379/" )
$redis = Redis.new(:host => uri.host, :port => uri.port, :password => uri.password)
When I console in to heroku I can do and get the following
irb(main):001:0> $redis
=> #<Redis client v3.0.1 for redis://carp.redistogo.com:9274/0>
Any help would be much appreciated. Thank you.
It looks like this:
GrnAuth::Application.config.session_store :redis_store, :server => APP_CONFIG['redis_server'], key: '_grn_session'
Should instead be this:
GrnAuth::Application.config.session_store :redis_store, :servers => APP_CONFIG['redis_server'], key: '_grn_session'
OK - I'm in way over my head here. I'm using:
- ruby-1.9.3-p0
- rails-3.1.3
- mail-2.3.0
- devise-1.5.3
Trying to turn on Devise's :confirmable option and start up smtp services in my app. As soon as I add /config/initializers/setup_mail.rb, add Devise's necessary columns in my DB, and the :confirmable attribute to my User model, I get a segmentation fault. It occurs right after a user signs up. Devise is trying to send out the confirmation email, causing the ruby smtp library to crash with the following:
... lib/ruby/1.9.1/net/smtp.rb:583: [BUG] Segmentation fault
The last entry in log/development.log:
Rendered devise/mailer/confirmation_instructions.html.erb (1.2ms)
My /config/initializers/setup_mail.rb file:
ActionMailer::Base.smtp_settings = {
:address => "smtp.gmail.com",
:port => 587,
:domain => "mydomain.com",
:user_name => "support#mydomain.com",
:password => "???????",
:authentication => "plain",
:enable_starttls_auto => true
}
ActionMailer::Base.default_url_options[:host] = "localhost:3000"
My config/environments/development.rb file has the following:
config.action_mailer.delivery_method = :smtp
config.action_mailer.raise_delivery_errors = true
config.action_mailer.default_url_options = { :host => 'localhost:3000' }
Here is the germane code from ruby/1.9.1/net/smtp.rb (line 583 is near the middle):
def ssl_socket(socket, context)
OpenSSL::SSL::SSLSocket.new socket, context
end
def tlsconnect(s)
verified = false
s = ssl_socket(s, #ssl_context)
logging "TLS connection started"
s.sync_close = true
s.connect # THIS IS LINE 583
if #ssl_context.verify_mode != OpenSSL::SSL::VERIFY_NONE
s.post_connection_check(#address)
end
verified = true
s
ensure
s.close unless verified
end
It looks like the segmentation fault occurs when smtp is trying to connect via a SSL socket connection ( s.connect ). In setup_mail.rb I've tried setting :enable_starttls_auto to both true and false. I don't get the segmentation fault when it is set to false but no email goes out, so that is useless.
I'm easily able to connect to gmail's smtp service by running this command from my Mac:
$ telnet smtp.gmail.com 587
Not sure where to go from here - any suggestions?
I had a very similar error (in net/http).
Doing this fixed it:
rvm pkg install openssl
rvm pkg install iconv
rvm pkg install readline
rvm reinstall 1.9.3 --with-iconv-dir=$rvm_path/usr --with-openssl-dir=$rvm_path/usr --with-readline-dir=$rvm_path/usr
The issue is with the way Rails interacts with OpenSSL. This post sums it up very well. http://www.22ideastreet.com/debug/smtp-rb14-bug-segmentation-fault/
The fix is to add this to your .bashrc/.zshrc/.bash_profile
export RUBYOPT="-ropenssl"
I've followd the "How I Test" screencast at RailsCasts, however I ran into a problem with spork
$ guard
Guard is now watching at '/Users/darth/projects/auth-before'
Starting Spork for Test::Unit & RSpec
Couldn't find a supported test framework that begins with 'testunit'
Supported test frameworks:
( ) Cucumber
(*) RSpec
Legend: ( ) - not detected in project (*) - detected
Using RSpec
Preloading Rails environment
Loading Spork.prefork block...
Spork is ready and listening on 8989!
ERROR: Could not start Spork server for Test::Unit & RSpec. Make sure you can use it manually first.
# here I get growl notification "Test::Unit & RSpec NOT started
Guard::RSpec is running, with RSpec 2!
Running all specs
Running tests with args ["--color", "--format", "progress", "--format", "Guard::RSpec::Formatter::NotificationRSpec", "--out", "/dev/null", "--require", "/Users/darth/.rvm/gems/ruby-1.9.2-p290/gems/guard-rspec-0.5.0/lib/guard/rspec/formatters/notification_rspec.rb", "spec"]...
.
Finished in 14.47 seconds
1 example, 0 failures
Done.
When I try to run spork in a separate terminal window, it doesn't help, as it gets killed instantly once I run guard
$ spork
Using RSpec
Preloading Rails environment
Loading Spork.prefork block...
Spork is ready and listening on 8989!
Killed: 9
If I just run spork and then try rspec . --drb, it works just fine. Here's link to gist with my Gemfile, Guardfile and spec_helper.rb
You should change:
guard 'spork', :cucumber_env => { 'RAILS_ENV' => 'test' },
:rspec_env => { 'RAILS_ENV' => 'test' },
:wait => 120
to:
guard 'spork', :rspec_env => { 'RAILS_ENV' => 'test' },
cucumber: false,
test_unit: false
This problem is actually caused by guard killing spork before it can even load, which is a problem on my slower MacBook pro.
The solution is to increase the wait time with :wait => 120 in the Guardfile, e.g.
guard 'spork', :cucumber_env => { 'RAILS_ENV' => 'test' },
:rspec_env => { 'RAILS_ENV' => 'test' }, :wait => 120
....
When I'm setting logging parameters to the Daemons (1.1.0) gem, how would I achieve similar behavior to this line?
logger = Logger.new('foo.log', 10, 1024000)
Daemon options:
options = {
:ARGV => ['start'],
:dir_mode => :normal,
:dir => log_dir,
:multiple => false,
:ontop => false
:mode => :exec,
:backtrace => true,
:log_output => true
}
Unfortunately the Daemons gem does not use Logger. It redirects STDOUT and STDERR directly to a file.
You can see the details of how the redirection works here:
https://github.com/ghazel/daemons/blob/master/lib/daemons/daemonize.rb#L241-261
Because of this, you will have to use something like logrotate and restart the daemon if you want to do log file rotation.
If this is not acceptable, I would suggest using Logger directly like you provided in the question.