5dc4b959fdedb2a9e22c71252fb14934a9a69361 Paul Dowman paul@pauldowman.com http://github.com/pauldowman/ec2onrails/commit/56ba4ad03315556d7ef0de12c6edef84b991103f 56ba4ad03315556d7ef0de12c6edef84b991103f

2008-05-19T07:39:57-07:00

Tagging release == 0.9.8 XXXX-XX-XX\n\n* ami-XXXX (32-bit) and ami-XXXX (64-bit) * This is another test release, still not intended to be production-ready. Look for version 1.0. Almost there!! * Incremental database backups instead of doing a full mysqldump each time (full dump is done once nightly). (Thanks to Markus Bengts for the bug fix) * Added monit monitoring daemon: monitors mysqld, apache, memcached, mongrels, system load and free drive space * Enabled postfix by default, mail is delivered to admin user (or use new ec2onrails[:admin_mail_forward_address] option in deploy.rb to forward admin mail to an email address) * Apache now supports SSL, set :ec2onrails[:enable_ssl] = true to enable and upload ssl cert & key (there is also a self-signed cert generated for each instance on first startup) * Base image now built with Eric Hammond's EC2 Ubuntu script: http://alestic.com/ (using svn revision 44) * Now installing latest rubygems (version 1.1.1) from source instead of using Ubuntu package * Apache now denies access to any file in a directory named .svn, to protect the subversion info of those who deploy with Capistrano "checkout" method * Added force-proxy-request-1.0 and proxy-nokeepalive for proxy requests to mongrel to fix occasional "proxy: error reading status line from remote server" error (mongrel doesn't support keep-alive anyway), see apache bug #37770 * Increased default memcache size to 128m, and memcache now listens on all interfaces for clustered use * Tuned MySQL setting defaults for better performance * restore_db_and_deploy task now runs migrations after restoring db * upgrade_packages task now does "safe upgrade" instead of "full upgrade" * Now requires Capistrano 2.3.0 * Bug fix: #20040: installing or upgrading packages no longer breaks - /etc/apt/sources.list is now correct * Bug fix: #19808: Capistrano database tasks don't try to load database config if there are no hosts in the "db" role (with :primary => true, slave roles not yet supported) * Bug fix: #19791: MySQL no longer fails after rebundling due to missing /mnt/mysql_data * Bug fix: #14555: Syslog messages about unacknowledged DHCP requests is no longer occurring * Bug fix: Database backup and restore scripts now exit if already backing up or restoring. * Bug fix: EC2 ephemeral store and swap volumes were not correctly mounted * Bug fix: Reload apache config when setting roles to find new app hosts. 6b7235c39c46dcfa3771081e75b064f7b40049c0 Paul Dowman paul@pauldowman.com