Engine Yard AppCloud

Jan 19

AppCloud v457 Deployed

• ENHANCEMENTS:  Ruby 1.9.2 users will find the database.yml generated with the mysql2 driver
• ENHANCEMENTS: Unicorn script issue where you saw ‘command not found’ is corrected and should no longer be displayed.

Jan 03

AppCloud v439 Deployed

• ENHANCEMENTS:  The Following updates have been made to the Unicorn control script.

1. Control script has  the path to include /data/<appname>/current/ey_bundler_binstubs which removed an OR condition which was confusing.
2. No longer tries to reload unicorn if it takes an excess of 30 seconds.
3. Changed the sleep periods started with a 30 second period if your application deploys in that short of a time, if not waits 60 more seconds longer.
4. If the deploy is ‘unsuccessful’ display the unicorn logs which can reveal why the deploy was ‘unsuccessful’  and then exit sanely not killing or touching unicorn any further.

• NOTE: If you use env.custom to enable a bundle exec unicorn you should revert those customizations as they are no longer required.

If you have any problems with the new unicorn control script please open an Support Ticket and or post on Our Community site if you do not have support.

Dec 16

S3 & EBS Volumes Are Slow

Though Amazon has not acknowledged it today, we are seeing many reports from our customers that S3 and EBS volumes are performing very slow today. This may affect your application’s performance.

Deploy Issues Have Been Fixed

We’ve tracked down the issue causing deploys to fail and fixed it.

Deploy problems still unresolved

We continue to experience deploy problems. We continue to work on the fix. We apologize for the inconveniences.

Deploy issue should be fixed

Deploy issue should now be resolved. If you continue to experience problems, please contact us.

Dashboard issues may cause problems deploying

Currently we are experiencing issues with our Dashboard. We are actively working to resolve them, and you are able to access it but there may be some problems deploying.

Dec 02

Failover Monitoring Problem

If a chef run occurred on any of your instances between November 29th and late December 1st, then the failover monitoring is most likely in a broken state. This is due to a bug that was introduced in version 0.2.5 of the Excon gem that’s used in failover monitoring. The gem was updated quickly and the fixed version is now being used, but you have to perform a full chef run to get the fix (a Rebuild if you’re using the CLI deployment method or a Deploy if you’re using the legacy deployment method).

If you encounter any trouble you think may be related to this, please contact support.

Dec 01

AppCloud v387 Deployed

• BUGFIX: Unicorn failing to reload for non-bundler users
• ENHANCEMENT: Unicorn now tries to bundle exec unicorn/unicorn_rails by default and if that is not available uses the system unicorn gem to run.  Additionally tries to resolve failed deploys in a reliable manner. 
• NOTE: If you use env.custom to enable a bundle exec unicorn you should revert those customizations as they are no longer required.
• UPDATE: Updated the Unicorn gem version to v1.1.5

Nov 09

AppCloud v371 Deployed

• BUGFIX: the dashboard is now firesheep-proof (SSL on all pages + Secure Connection-only cookies).