Release Notes

v2.10.6 (June 10, 2013)

• [98980] New kato patch command
• [96962] New Google Apps authentication strategy for AOK
• [96395] Upgrade ActivePython distribute version to the latest
• [98481] Fix a postgresql readiness issue on 'kato data export'
• [98785] Upgrade Nginx to v1.2.9
• [98626] Fixed 2.6.7 to 2.10 cluster data import error
• [98632] pyopenssl now installable via PyPM without an ActivePython Business Edition license
• [98602] Fixed orphaned erlang beam processes on rabbitmq service deletion
• [98636] Fixed filesystem service capacity limits
• [98626] Autoscaling: Fixed ArgumentError in get_datastore (1 for 2)
• [98681] Correct ownership of /home/stackato/.ssh/known_hosts
• [98715] cloud_controller X-Accel-Mapping header missing
• [98722] Locked down redis_server port from container access
• [98735] Management Console: Overview now shows number of nodes in cluster
• [99133] App Store: re-enabled Services filter
• [98875] Fixed cloud controller error when increasing app instances in Management Console
• [98884] Cloud controller now sends router.unregister NATS message on shutdown or restart
• [98928] Fixed incorrect status codes in router2g access log
• [98942] Hooks with grails (java_web) framework fixed
• [98590] Fixed kato status role reporting overlap
• [98972] Fixed MySQL service metadata import
• [98985] Less aggressive DEA autoscaling
• [98997] Fixed potentially blocking operation during retry of TCP drains
• [99006] kato op upstream proxy ... now correctly modifies the LXC template
• [99025] Longer (and configurable) timeout for kato op regenerate postgresql
• [99040] Management Console now un-maps URLs correctly
• [98966,99012] fixed broken postgresql first boot task
• [99145] Fixed an issue where services may not be deleted from the web console
• [99153] Fixed Harbor service issues with UDP backend health checks
• [99220] Fix for cloud-init Stackato 'rename' task failing
• [99230] Fixed cloud-init failing to install custom apt-packages
• [98997] Fixed potentially blocking operation (up to 10s) in drain:stop
• Fence/DEA (app container management):
  • [98979,98990] More accurate disk usage monitoring
  • [99100] Fixed a DEA memory leak with droplet management
  • [98749] Fixed FATAL -- Can't create DEA pid file: Process already running
  • [98700] Fixed traceback on DEA restart (parse_info!': undefined method `[]')
  • [98688] Fixed Filesystem::node: No such file or directory - /var/vcap/sys/run/fence.sock error
  • [97619] Improved the accuracy of (accounted) container memory usage
  • [99147] Ensure container creation failure does not crash fence
  • [99204] Fixed "No such file or directory" ENOENT error on DEA
  • [98934] Container cleanup fix
  • [99202] Fixed "can't add new key into hash during iteration" warning on container cleanup
• stackato client updated to 1.7.4
  • [97359] New default behavior for stackato update preserves previously set environment variables. New --env-mode replace option enables the old behavior of removing or resetting environent variables.
  • [96962] New token command for interactive token-based authentication with external SSO systems.

v2.10.4 (April 5, 2013)

• [97520] Oracle DB add-on with stackato dbshell support
  • ORACLE_URL and JDBC_ORACLE_URL env vars if Oracle service present
  • Spring auto-configuration of Oracle DB binding
• Improvements to Logyard:
  • [96008] Fix logyard crash and uncleaned drains sometimes triggered by kato log tail
  • [98244] Delete app drains on stackato delete
  • [97614] Improve apptail reliability
  • [98279, 98354] Configurable finite retry on drains
  • [98169] Configurable maximum limit on user drains
  • [98326] Support for named custom format for use in drain URIs
  • [97856] apptail: handle NATS disconnection
  • [98280] Fix retrying behavior to reset retry delays after some point
  • Simplified formatting of log records in logyard.log
  • [98354] Reduced the frequency of retry warnings (esp. for Cloud Events)
  • Stability improvements [98091, 96337, 98342, 98243, 98240, 98423]
• Core runtime updates for stability and security [97667, 97978, 97857, 97551]
  • [98515] Upgraded PostgreSQL to 9.1.9 to address CVE-2013-1899 (major), CVE-2013-1900, and CVE-2013-1901
  • [97645] Improved cookie handling of token
  • [97649] Prevent REST API cross site request forgery
  • [97646] Prevent XSS vulnerability through App Store yaml file
  • [98256] Address CVE 2013-1857 in Rails
  • [98521] Reduce controller default memory limit to 50% (from 70%) before it is restarted
  • [98050] Updated doozerd
• Node.js
  • [97557] Updated node08 runtime (0.8.22)
  • [98068] Added node010 runtime (0.10.1)
• Perl:
  • upgrade to uWSGI 1.4.9
  • updated ActivePerl-5.14 runtime (5.14.4.1405)
  • added ActivePerl-5.16 runtime (5.16.3.1603)
  • latest App::cpanminus supports installation via URLs and from git
  • added support for cpanfile to specify dependencies
  • added experimental support for Carton, the Perl module dependency manager
• Python: upgrade to uWSGI 1.4.9
• Ruby:
  • [98316] Update Ruby autoconfiguration library for MongoDB
  • [98223] Improve handling of stdsync usage
  • [98113] Updated Ruby buildpack
• Router improvements
  • [97806] Implement NATS connection error handling and retry logic
  • [98292, 98252] Fix possible JSON truncation in large responses
  • [98403] Correctly drop lost controller nodes from routing table
  • Additional fixes [96790, 98062, 98029, 97882]
• kato updates:
  • [95989] kato users has moved to kato data users
  • [97251] Added kato node list to list all nodes and which roles are running on them
  • [98137] kato tail skips logyard INFO records unless --raw is passed.
  • [98107] Correct kato node rename to handle existing pushed apps when using mdns
  • [97934] Prevent kato data export exception if memcached services was never used
  • Additional bug fixes [98067, 96792, 97597, 97769, 97808, 97810, 97753, 97392, 97761, 97752, 97809, 97561, 98009, 96795, 98384, 98272, 98481, 98402]
• [98281] Fix an issue with stopping apps during DEA shutdown.
• [96175] Fix a regression in stackato files logs/staging.log not returning staging.log
• [97895] Fix yajl load error in wait_for_fs.rb (when filesystem service is used)
• [98282] Speed up ssh connections to system
• Web console improvements [97736, 97702, 95765, 97701, 96121, 97777, 98144, 97699, 98140, 97996, 97882]
  • [98254] Display Logyard configuration settings
  • [98042] Fix memory leak on long-open pages polling cloud events
  • [98127] Instant refresh of in-page modified components
• [97025, 98259] Improve DEA internal connectivity speed
• [98281] Improve DEA app cleanup on shutdown
• [93626] Enforce random postgres password at first boot
• [96007] Fix JBoss database support error when no services are defined
• [97731] Reduce long polling calls from the web console
• [97772] Correct ability to do first user setup after kato factory reset
• [97993] Made log rotation more strict (rotate any log in /s/logs/ over 10MB)
• [91120, 98299, 97973] Usernames are now stored in a case insensitive manner
• [98182] Updated Mono framework add-on to 3.0 Beta (Mono version 2.11)
• [98312] Update user authentication check to return compatibility to vcap_java_client
• [97322] cloud-init updates
• [97253] Change of "cloud_controller" role to "controller", new "primary" role
• [98314] Restrict kernel messages to tty console to errors only
• [98289] Improved console tty startup process to show progressive start state
• [98288] Update MongoDB to 2.4.1
• [97322] Improved cloud-init support to allow some kato commands to be called deferred
• [98523] Make MySQL service STORAGE_QUOTA_INTERVAL and KEEP_ALIVE_INTERVAL configurable, move quota check to every 30s default
• [97760] Properly handle timeout errors when checking DB quotas
• [97619] Updated app instance memory calculation to use container group stats
• [98550] Set container memory limit exactly at app instance memory request (was overallocating 12.5%)
• Documentation updates
• stackato client updated to 1.7.2
  • [98187] New stackato admin grant/revoke/list
  • [98145] Fixes to some false negative key warnings
  • [98221] Fixed multi-app push from stackato.yml
  • [98045] Correct limits management between users and groups
  • [96810] Correct handling exploded .war directories
  • Additional bug fixes [98238, 96930, 98181, 92751, 98089, 97958, 97906, 97520]

V2.8 (February 26, 2013)

• App Store improvements
  • Now a separate component.
  • Full streaming of application deployment output to the Management Console
  • [96401] App Store output now available in app log stream (stackato logs)
  • [97622] App Store uses upstream proxy info (e.g. provided by kato op upstream_proxy set).
• New Harbor port service
  • Allows external ports to be forwarded directly to your app
  • Supports TCP, UDP and dual TCP/UDP ports.
  • Supports multiple app instances (TCP only)
  • [95358] Supports use of custom and multiple SSL certs with your application. For apps that use frameworks, harbor provides a built-in HTTPS server with SPDY support for the custom certificate(s).
  • Passive health checking of unavailable backends or unresponsive apps.
• Fence (fully non-blocking container management):
  • Less memory usage, faster DEA/stager performance
  • Faster container startup
  • Addition of per-app enforced disk limits
  • Improved security via Unix user-level compartmentalization in the kernel (each container now has its own Unix user)
  • Support for any IP range for LXC containers (configurable)
  • Removing (deprecating) lxctrl processes (less memory use)
• AOK (user authentication):
  • Optional external user authentication component
  • LDAP authentication support
• [97274] PostgreSQL: HSTORE (key/value storage) support in the PostgreSQL service.
• Redis: upgrade to v2.6 (release notes)
• Python: upgrade to uWSGI 1.4.5
• Perl: upgrade to uWSGI 1.4.5
• Java: JPDA debugging for java_web applications via Harbor port service.
• Go: upgrade to the latest Go buildpack
• Logyard updates
  • [96620]: Drains are now retry infinitely if disconnected.
  • [96338]: Added new fields to the JSON of application log stream: AppID (application ID), AppName (application Name), NodeID (host DEA IP of the instance)
  • Added a new file:/// drain for debugging purposes.
• [96493,96498] UI improvements for non-admin users in Management Console.
• [95986] New 'app-dir' key in stackato.yml to support deploying from a specific sub-directory.
• [97013] No port lookup for app host in incoming HTTP header.
• [95902] Management Console: Warn when deleting the last role on a node.
• [96729] Admin-configurable web console login expiration
• [96700,96653] Rails security fixes.
• [96677] mDNS backend: avahi-daemon removed and avahi_announcer replaced with avahi_publisher and app_mdns.
• [96225] Manual SSH key configuration on systems with separate Router, Cloud Controller, or Load Balancer nodes is no longer required during setup.
• [96696] Allow simultaneous 'kato tail' sessions.
• [96932] Fix certain /files routes e.g 'stackato files' from not being routed correctly.
• [95576] Prevent CC from being blocked during a vSphere scaling operation
• [96708] Major changes to kato commands and command naming:
  • "kato start|stop|restart" now manipulates roles, not processes.
  • "kato enable|disable" is now "kato role add|remove"
  • "kato attach" is now "kato node attach"
  • "kato setup micro" is now "kato node setup micro", "kato setup core" is now "kato node setup core" and "kato setup load_balancer" is now "kato node setup load_balancer"
  • "kato setup controller" and "kato setup node" have been retired. Use "kato node setup core" and "kato node attach" instead.
  • Change to how "kato drain add" options are passed. Dropped --option key=value --option key=value in favor of a just key=value key=value
  • For "kato role add|remove" (previously kato enable|disable), --all-but and --only are now boolean flags that act on the list of roles given. Therefore "kato enable --all-but dea,prealloc" becomes "kato role add --all-but dea prealloc"
  • "kato role remove" no longer supports the flag --no-stop
  • "kato rename" is now "kato node rename"
  • Added "kato role list" to list all roles and number of nodes they are enabled on.
  • "kato config" now takes action keywords "set", "get", "del", "push" and "pop". This is not backwards compatible. Example is "kato config set <component> <key-path> <value>".
  • "kato tail" is now "kato log tail"
  • "kato drain" is now "kato log drain"
  • "kato process start|stop|restart" can now be used to manage processes, though managing roles via "kato start|stop|restart" is recommended.
  • "kato ready", for checking the ready status of processes, is now "kato process ready"
  • Add "--node" to "kato process ready" to check status or process on any node.
  • "kato config" default output is YAML
  • "kato debug lsdoozer" has been removed. Use "kato config get --flat" instead.
  • The previous "kato status" functionality that showed just a list of processes and their statuses can now be found under "kato process list"
  • "kato import" is now "kato data import".
  • "kato export" is now "kato data export".
• stackato client updated to 1.7.0
  • [97399] Added '-d' option to enable JPDA debugging via Harbor service.

v2.6.7 (December 19, 2012)

• [96341]: Fix --format not being respected in kato drain add
• [96225] Manual SSH key configuration on systems with separate Router, Cloud Controller, or Load Balancer nodes is no longer required during setup.
• [94390] Added Node.js v0.8.14 as a dual runtime along with v.0.6.20
• [96231] Java 7 has been added as a new runtime.
• stackato client updated to 1.6.2
  • [95509] References to 'stackato' in renamed client help fixed.
  • [96386] Fixed output problems with 'stats' command.
  • [96427] Fixed problem with '--no-tail' option.
  • [96429] Fixed "Ambiguous runtime" error with 'standalone' framework.
  • [96438] Added option '--token' for direct specification of the auth token.

v2.6 (November 27, 2012)

• Log aggregation/forwarding updates (Logyard)
  • Application log streams
  • Redesigned to run on large-scale clusters with no single point of failure or inter-node traffic.
  • Better log forwarding via drains (tcp, udp, redis).
  • Added Splunk support (based on drains).
  • Key deployment events made available in the application log stream.
  • Improved documentation.
• Cloud-Events
  • Cluster support - now includes events from all nodes in the cluster.
  • Forward events via Logyard drains.
  • Build reports on top of cloud events via third party services.
• Kato import/export changes - new next generation cluster migration tool
  • kato import/export uses the cluster-aware back end formerly accessed with kato migrate (deprecated)
  • Bug fixes: [96065, 96072, 96073]
• Client updated to 1.6.1
  • Application log forwarding setup via stackato drains add (user-level)
• New kato history command shows a history of administrative kato commands ran across the cluster.
• Removed Command: kato ls in favor of kato status --all
• [93012] Allow multiple staging nodes
• [96026] Create "contrib" plugin frameworks directory
• [95935] Fixed misleading 'Cannot connect to doozerd' warning in various cluster situations.
• [96215] Security fix for app store installs
• [96114] New dedicated Nginx instance for handling Cloud Controller file uploads
• [94258] Configurable data services storage location
• [95703] Fixed headers being prematurely sent on error pages with router2g
• [96191] Upgrade to Ruby 1.9.3-p327
• [96255] Fix issues with kato node rename when /etc/hostname is empty
• Additional documentation for new features, plus edits and minor corrections

Release v2.4 (October 29, 2012)

• Improved application logging:
  • Logs are persisted between application restarts
  • stackato logs runs much faster, returns logs from all instances, and returns logs from stopped/crashed instances
  • stackato logs --follow tails logs from all instances
• Improved system logging (BETA):
  • Aggregate stackato system logs from all nodes in the cluster
  • Output plugins to stream system logs to Loggly, papertrail, and others
• Import and export Users and Groups via web console (.csv format)
• Simplified web console theming, with improved documentation
• Upgraded Apache TomEE Plus from v 1.0.0 to v1.5.0. In addition to the support of new TomEE features as listed on the TomEE site, this upgrade addresses additional bugs:
  • [95426] Added Hibernate
  • [95082] Fixed Grails apps
  • [95425] Fixed test failure (Java apps with mysql services)
  • Other application compatibility issues
• New Command (BETA): kato migrate - next generation cluster live migration tool
• New Command: kato op static_ip - intuitive tool for static-ip configuration
• New Command: kato relocate ... - move key shared data to a new location (e.g. persistent storage)
• Client updated to 1.5
  • [95702] Allow uploading very large applications (upload limits still imposed by server)
  • [92160] New subcommand scp to move data to/from app instances
  • Support for enhanced logging features
• Client installation no longer available via pypm
• MongoDB upgraded from v1.8.1 to v2.0.7
• Node.js upgraded from v0.6.18 to v0.6.20
• Ruby upgraded from 1.9.3-p125 to 1.9.3-p194 (with security patches from p286)
• [95291] Fixed incompatibility with Eclipse Cloud Foundry plug-in and Spring Tool Suite
• [94769] Allow non-web background processes with the "standalone" framework
• [94734] Support non-web background processes for all other frameworks
• [95491] Show DEA host node IP in application Details page for admins
• [94076] Auto-Complete for adding users to groups
• Experimental 'load_balancer' feature available via kato setup (BETA - not for production use)
• Router2g improvements (BETA):
  • SPDY support on all HTTPS connections
  • Improved logging of events
  • Support multiple cloud controllers
  • [95280] Fixed not setting content-length for Django applications
• Tuned system logging behavior for some components (DEBUG->DEBUG2)
• Ability to deselect App Store URLs without deleting them
• The App Store can now be disabled (removed from the side menu)
• Allow auto-starting of apps deployed from the App Store
• Documentation for new features, plus edits and minor corrections

Release v2.2.3 (September 28, 2012)

• [95661] Revert to older version of supervisord to fix an issue with cluster setup

Release v2.2.2 (September 20, 2012)

• Integrated Ruby buildpack. When using framework: buildpack in stackato.yml for ruby apps, BUILDPACK_URL can be omitted.
• Java: Correct copy_source_files for non java_* frameworks
• Java: Improved access to logs
• Java: Add logs for debugging EAR JEE applications
• [95287] Improved memory usage calculation in forked Apache process apps
• [95290] Tuned default Apache settings for app instances
• [95458] Prevent doozer "TOO_LATE" warning
• [95265] Handling client errors in App Store deployments more gracefully
• [95314] App Store checks if application name is already in use
• Client updated to 1.4.5
  • Fix Windows Unicode pathname handling
  • [95485] Tail stderr.log instead of startup.log on push in slow deployments. Latter file is not used by all frameworks.
• Documentation updates:
  • New Relic integration docs updated
  • Updated cluster docs to use the correct form of kato attach and switched from kato setup to kato attach in autobecome
  • Added pyramid-default sample to Python docs
  • Updated Ruby deployment information
  • Updated autoscaling section with EC2 URL
  • Correct use of the Authorization HTTP header in client APIs

Release v2.2 (August 15, 2012)

Development

• Numerous additions and changes in kato server administration commands:
  • Most of kato rewritten in Ruby
  • kato node {remove|migrate} to manage outdated nodes or changing core IP
  • kato ready to determine ready status of underlying components
  • kato inspect to check for common configuration issues
  • kato setup core to replace kato setup controller
  • kato attach to replace kato setup node
  • kato enable|disable data-services
  • [94947] Manage cloud_controller as independent role to support multiple instances
  • Improve help output and error handling for incorrect arugments
• [94844] stackato run dbexport: exports mysql and psql databases
• [94657] Changed the -n alias for kato setup -hostname to -a
• [94357] Filesystem service split into filesystem_gateway and filesystem_node to make multiple filesystem service nodes possible
• [94905] Adjust VCAP_SERVICES service key to use unversioned service name (version remains in label key) to bring parity to client service creation
• [94783] Experimental router2g with WebSockets support
• [94744] Updated doozerd binary, fixed verbose doozerd.log logging affecting disk i/o
• [94828] Properly url encode console links for user names with special characters
• [94501] Extract droplets asynchronously on DEA
• [92457] Enable kato export via web console
• [94638] Expose Allow Overprovisioning flag for all services
• [94823] Switch App Store format from json to yaml and add tags field
• [95204] Use core stackato client for App Store installation and make asynchronous [94834]
• [95130] Auto redirect to logout when old or bad credentials are passed to deep linked page
• [94734] Fix handling of file system links when viewed in web console
• [95090] Defer re-staging of stopped apps when application environment is modified to next start request.
• [95049] Ensure log rotation of cloud controller rails log
• [94933] Improve handling of postgres restart in kato import
• [95151] Fix auto-incrementing of uniq keys in kato import
• [95156] Enable apt-get blacklist of some core packages to ensure system stability in system apt-get update
• [93847] Set max_droplet_size via kato op max_client_upload
• [95211] Run pre-running hooks after env generation during deployment to ensure database migrations are finished before the user commands execute
• [95082] Fix Grails incompatibility with TomEE due to conflicting validation jar
• [95028] Add JBoss framework JDBC service handling
• Updated Stackato client to 1.4.4:
  • [95197] Improved logout command to allow selective target logout
  • Fixed grails [95112] and aspdotnet [95122] framework auto-detection
  • Enhanced variable resolution in stackato manifest to resolve variables in the key portion of the yaml
  • Added EAR archive detection for JavaEE framework
  • Merge command line options from push into the stackato manifest values
  • [94852] Compare http headers case insensitively

Documentation

• Clarify usage of dnsmasq
• [93980] Document how to configure DNS to support access to an application URL
• Added .NET deployment section
• Improvements in:
  • client/core API docs
  • language deployment (java, VCAP_SERVICES, DATABASE_URL)
  • stackato.yml docs (variable subsitution etc.)
  • cluster & auto-scaling docs. (kato setup core)
• App-store docs rewrite. (JSON -> YAML based store, simplified)
• Best-practices docs rewrite. (migrations, export, import, dbshell...)
• Moved FAQ page to http://community.activestate.com/faq-list?tid=453
• Deprecated Kato setup controller/node. See kato command reference.

Release v2.0.7 (Jul 19, 2012)

• [94813] Fix for kato import failure when migrating from v1.x to v2.x
• [94737] Fix for potential sudo errors when sudo is enabled in the containers
• [94716] Fix for special characters causing issues in user email address
• [94740] Fix for NATS timeouts in the Stager
• [94578] Fix for Stackato usage reporting accuracy
• [94847] Fix for more precise memory usage calculation in deployed applications.
  • Improved container cleanup managment
• [94836] Optimisations for DEA app queuing after cold startups of Stackato.
  • Configurable simultaneous app starts per DEA
  • Configurable maxmium prealloc queue length
• [94812] Improved first setup workflow for statically configured IP environments.
• Update Stackato client to v1.4.2
• Various documentation improvements and updates

Release v2.0 (Jul 10, 2012)

• Better cluster management
  • kato, a better replacement for stackato-admin
  • Central multi-node configuration through doozer
  • Better process management through supervisord
  • CLI to edit configuration on all nodes: kato config
  • Fix: make nats-server run only on the controller node
  • Fix: automatically restart avahi-announcer and redis-server if they crash
• Merged with latest Cloud Foundry sources
  • New Memcached as a service
  • Router v2 upgrade
  • Configurable restart behaviour for crashing apps
  • New Java Play framework, also supported via Buildpacks
• Allow stager to be run on different node
• Avahi / mDNS now compatible with a cluster setup
• Improved kato tail; now includes non-vcap components such as nginx, postgres, redis.
• Base OS upgraded to Precise Pangolin (12.04)
• Database service versions upgrades
  • Postgresql upgraded to v9.1 (-> v8.4)
  • MySQL upgraded to v5.5 (-> v5.1)
  • Redis upgraded to v2.4 (-> 2.2)
• Python
  • Custom pypm/pip options via $PYPM_OPTS and $PIP_OPTS environment variables
• Scala support via Buildpack
• Java EE 6 support
  • New java_ee framework via Apache TomEE
• Integrated support for Iron Foundry .NET runtime
• Runtime upgrades
  • Node.js v0.6.18
  • PHP 5.3.10
• Stability bug fixes.

Release v1.2.6 ( May 15, 2012)

• Fix for an issue using the stackato-admin update-services-ip command
• Fix for missing Komodo debugging bits
• Fix for a groups issue when migrating from Stackato v1.0 -> v1.2
• Minor documentation updates

Release v1.2 (Apr 27, 2012)

• Upgraded Runtimes:
  • Ruby 1.9 upgraded to 1.9.3-p125 (from 1.9.2-p290)
  • Node.js upgraded to v0.6.15 (from v0.6.10)
• Updates to Stackato client.
• Heroku buildpack support, with built-in buildpacks for Clojure, Go, Java, Play! and Python
• stackato push|update will now tail the staging.log file simultaenously (like Heroku).
• Support for user groups
• Support for per-user and per-group limits
• Improved dbshell: stackato dbshell is deprecated in favour of stackato run dbshell. The later works on cluster environment, and doesn't require local database clients to be installed.
• Added STACKATO_SERVICES as an alternative to VCAP_SERVICES; PORT as an alias for VCAP_APP_PORT.
• Added new service environment variables containing the respective connection URIs: DATABASE_URL, MYSQL_URL, POSTGRESQL_URL, REDIS_URL, MONGODB_URL, RABBITMQ_URL
• Allow SSH'ing to crashed app instances
• Optional support for caching staging files for faster stackato update
• Disable truncation of vcap logs on component start. This preserves the log records when a vcap process is restarted.
• Added support for OpenStack auto-scaling
• Use geo-location aware debian mirrors in application containers for faster apt-get calls
• Improved the firstboot configuration process, eliminate a timeout bug in some cases.
• Improved, faster VM boot process (Moved to Upstart).
• VM now uses NTP for time synchronization
• Fix for supporting externally mapped URI's on applications
• Fix for Avahi broadcasting the wrong inteface address
• Add Go language support
• Add PyPy support (external buildpack required)
• Add JRuby support (external buildpack required)
• Python:
  • Allow pip to use mirrors (--use-mirrors) when installing packages
  • Upgrade to uWSGI 1.1.2
  • uWSGI now supports running Python 3.2 applications
• Perl
  • Upgrade to uWSGI 1.1.2
• Node.js:
  • All packages installed from NPM are now automatically added to the $PATH, meaning sudo priveleges for 'npm install -g' are no longer required.
• Clojure:
  • Upgrade to Leiningen 1.7; switch to Heroku Buildpack
• Automated DBGP Debugging
  • Pushing an app with --stackato-debug <host>:<port> enables Django, Rails, and Sinatra get requests to be debugged in DBGP-supporting apps like Komodo IDE with a STACKATO_DEBUG=1 parameter, and enables PHP debugging with the XDEBUG_SESSION_START=1 and XDEBUG_SESSION_STOP=1 parameters.

Release v1.0.6 (Mar 08, 2012)

• Fix an issue for Filesystem services not unprovisioning correctly
• Fix an issue with LXC container memory management causing a potential system lockup

Release v1.0 (Feb 28, 2012)

• Revamped Stackato Management Console
  • Increased the administrators control of the overall system
  • Ability for users to control their own applications
  • Core system server performance graphs
  • Improved cloud events logging
• New persistent file system service
• Support for the Cloud Foundry (vmc) manifest.yml
• Added "Generic framework" to deploy applications with a arbitrary start procedure
• Support for dynamically pre-allocating application containers.
• make app environment variables available to staging, run commands, ssh and cron.
• stackato-admin report: wrap logs and other system information for future diagnosis of a problem.
• The applications timezone is now inherited from the Stackato VM host
• $app_http_proxy is now available to specify a application specific $http_proxy
  • The new stackato-admin upstream-proxy command helps setting this variable, check the docs or stackato-admin help for more information.
• Moved vcap logs in /tmp/vcap-run/ to $HOME/stackato/logs/
  • These logs are now compressed and rotated daily
• show the actual staging log, instead of Error 310: Staging failed, upon push / update failures
• allow top-level dotfiles during staging
• stackato-admin become <node> now allows renaming operations on the VM, allowing you to speficy the FQDN and local hostname.
  • The shorthand all role has also been added to stackato-admin become to assign all the available roles.
• Bug fix for sporadic high load averages on DEA nodes
• NewRelic Server Monitoring bits are now preloaded onto the VM.
  • Get Started
• Implemented fallback console on the virtual TTY1 for unsupported resolutions.
• The TTY application now has fundamental error reporting built in.
• git, mercurial and subversion are now made available in application instances.
• Staging support for manifest.yml
• Support for a external MySQL Database (e.g existing cluster)
• Python:
  • Run pip with requirements.txt, as that is the most common filename used by Python projects; if only requirements.pip exists, run pip with that file. For PyPM, use stackato.yml or requirements-pypm.txt.
  • Fix: make "-e" (source installs) in pip requirements.txt work
• Clojure:
  • Automatically run lein deps to download dependencies
• Node.js:
  • Upgrade to v0.6.10
• Ruby 1.8:
  • Upgrade to patchlevel p-357

Beta Release 0.8.2 (Jan 17, 2012)

• New web-based management console for administrative control of Stackato
• Merge with Cloud Foundry upstream as of Dec 21, 2011
• Application Staging now happens in LXC; thus more secure.
• Support for SSH'ing into application instances. (experimental)
• Improved and reliable stackato run based on SSH.
• Support for schedule tasks through cron (experimental)
• OpenStack build support
• Python:
  • Fix PYTHONPATH to contain correct value in the "run" command
  • Use PyPI mirrors during pip staging
• Java:
  • Move from Sun/Oracle Java packages to Open JDK/JRE packages (why?)
• Node.js:
  • Upgraded to version 0.6.7
• Clojure:
  • Leningen upgraded to version 1.6.2
• VM:
  • Added dynamic first boot configuration process to randomise the hostname and SSL certificates.
  • New console application running on tty1
  • Microcloud HTTPS/SSL certificate now supports the TLD and all subdomains
  • Udev networking rules have been disabled

Beta Release 0.6.0 (Nov 03, 2011)

• Stackato admin dashboard (start with stackato-admin start dashboard)
• Faster app deployment
• Set user's home directory to correct location within LXC containers
• Perl:
  • Remove Perl 5.12 and make Perl 5.14 the default
  • Store uWSGI command line in a $STACKATO_UWSGI environment variable.
  • Store the default Perl application startup command in a $PROCESSES_WEB environment variable (TBD for other frameworks)
• Python:
  • uWSGI for Python 3 is now officially supported
  • Store uWSGI command line in a $STACKATO_UWSGI environment variable.
  • Store the default Python application startup command in a $PROCESSES_WEB environment variable (TBD for other frameworks)
  • Add virtualenv's bin/ directory to $PATH before starting the application

Beta release 0.5.2 (Oct 20, 2011)

• Upgrade ActivePython to 3.2.2.3
• Upgrade Node.js to v0.4.12
• Upgrade Ruby 1.8 to patch level 290
• Additional LXC fixes for non starting apps
• Fix for broken clojure deployments

Beta release 0.5.1 (Oct 3, 2011)

• LXC bug fixes and improvements (Error 306, etc.)
• Speed up PyPM/PPM/cpan/pip by caching file downloads from the internet
• Support for running custom hook scripts (post staging and pre running)
• Upgrade to uWSGI 0.9.8.5 (LTS)
• PHP support
• RabbitMQ support
• Python:
  • Properly show pip errors (during push/update/log) in staging.log
  • Add the application directory to $PYTHONPATH
  • uWSGI: remove --catch-exceptions by default
  • uWSGI: allow applications to spawn threads
  • uWSGI: don't spawn an unnecessary additional server process
• Java:
  • Add :port back to VCAP_APPLICATION for the cloudfoundry runtime to work
• Server process monitoring (via monit)
• Known issues:
  • Clojure deployments are currently unavailable due to a change in the staging process.

PDP4 (Aug 25, 2011)

• Instances are isolated using LXC containers on DEAs
• Clojure support
• General improvements to error reporting in staging.log
• Server now returns the git version info to client
• Documentation is hosted locally at http://docs.stackato.local
• stackato run
  • Include app environment added via stackato env
  • Wrap commands with quotes when necessary
  • Provide a dbshell command to invoke database clients
  • Run command no longer blocks the EventMachine reactor
• Python
  • Don't silently ignore packages missing in pypm repository
  • Deprecate virtualenv in favor of PEP 370 (faster deployments)
  • Allow invocation of scripts installed by package dependencies
  • Can now install packages with ./include dir. (eg: greenlet)
  • Drop 'Generic Python' framework
• Ruby
  • Disable gem caching as it doesn't work with secure mode
• MySQL
  • Use UTF8 as the default database charset

VM

• Upgrade to ActivePython 3.2.1.2
• Upgrade to NodeJS 0.4.10

PDP3 (July 13, 2011)

• Support for PostgreSQL service
• Fix random 306 errors when push/update fails
• stackato run:
  • Fix 306 errors after ~11 seconds
  • Suppress traceback for ENOENT
  • Kill the process after timeout
  • Make PPM area directory available
  • Cleanup $PATH; add all runtimes to $PATH
• Python
  • Note pypm/pip failures in staging.log
• Node
  • Auto install dependencies using npm install

VM

• Upgrade to ActivePython 2.7.2.5

PDP2 (June 22, 2011)

• stackato run:

  • Put runtime's path in front of $PATH

  • Close STDIN descriptor to prevent hanging processes

  • Support --timeout option (default = 1 minute)

  • Ruby: set GEM_{HOME,PATH} environment variables so dependent

    binaries (eg: rake) can be run

• Python

  • pip and pypm can both be used to install Python

    dependencies

  • Python 3.x support for 'Generic Python' framework

• Perl

  • Perl applications must have a toplevel app.psgi file
  • Perl 5.14 support under the "perl514" runtime name
  • Perl modules can be installed with cpanm via Makefile.PL
  • new Perl logfiles: logs/staging.log and logs/ppm4.log

• Ruby:

  • Fix an issue with linking 1.9 gems with ruby1.8

VM

• Support multiple uWSGI binaries (PSGI, Python 2.7, 3.2)
• No more uWSGI zombies
• Installer creates stackato service script in /etc/init.d
• stackato-admin controls the avahi-announcer's lifecycle (no more need to start/stop it separately)
• Stackato client is installed on VM (necessary for multi-vm setups)
• Resource-efficient avahi-announcer
• avahi-announcer announce the apps names, not the mapped URLs
• stackato-admin can manage admin users

Internal

• Remove the vcap/installer/remote after an install
• stackato-admin uses yaml parser instead of grep
• Install script is now called install.sh and accepts parameters
• Do not regenerate the mysql password on restarting ./install
• Don't pull uWSGI directly from mercurial repo
• Allow git branch argument for ./install

PDP1 (June 15, 2011)