RELEASE

    ___                _____            __    
   /   |  ____  ____  / ___/_________  / /___ 
  / /| | / __ \/ __ \ \__ \/ ___/ __ `/ // _ \
 / ___ |/ /_/ / /_/ /___/ / /__/ /_/ / //  __/
/_/  |_/ .___/ .___//____/\___/\__,_/_/ \___/ 
      /_/   /_/                               

AppScale version 2.7.1, released February, 2016
Highlights of features/bugs in this release:
- A bug in get_app_data that affected upgrades was fixed
- AppScale can be built on Debian Wheezy
- AppScale can be built on a Raspberry Pi 2
- Python applications can now use the vendor module
- Datastore debug logs are more useful
- There are fewer ZooKeeper connection errors

AppScale version 2.7.0, released February, 2016
Highlights of features/bugs in this release:
- Fixed bug in choosing app ports via lsof
- Fixed Vagrant FastStart to pick the correct IP
- Added hourly status log in AppController
- Handle UpdateIndex requests to the Datastore
- Removed hardcoded refs to AppScale home dir
- AppScale now runs on Ubuntu 14.04 LTS (Trusty Tahr)
- Have Monit check app servers via the port
- Increase Java URLFetch API size limit to 33MB
- Clean up expired successful transactions
- Fixed bug in monitoring Nginx
- Updated URLFetch stub to handle socket timeouts
- Keep about 1GB of app logs
- Replaced nc with logger for centralized app log
- Made start services idempotent
- Have Zookeeper autopurge run more often
- Use Monit to stop all running services during appscale down
- Fixed bug that was spawning additional cloud instances
- Better AppController restore flow
- Fixed bug that was causing an app to be disabled
- Assigned application ports will now persist through down/up
- Use Xenial's version of kazoo
- Fixed AppController's stop command
- Removed the use of root RSA keys
- Have the AppManager set up AppServer routing
- Create separate logrotate script per application

AppScale version 2.6.0, released December, 2015
Highlights of features/bugs in this release:
- Fixed bug that was preventing large blob uploads for Java apps
- Avoid unnecessary data decoding when reading from the Datastore
- Configure Monit on the server side
- Better handling of crash logs on the Tools side
- Better handling of health checks when overriden by the app
- AppScale build in Docker container
- Fixed race condition when allocating application ports
- Faststart ability on all supported infrastructures
- Java SDK upgrade to 1.8.4
- Fixed bug with quotes for Java cron jobs
- Limited the number of Zookeeper logs
- Fixed permissions when untarring app code
- Fixed bug in Groomer Monit configuration
- Various Recovery improvements

AppScale version 2.5.0, released November, 2015
Highlights of features/bugs in this release:
- Misc AppController fixes related to port collisions
- Handle reserved delimiters in property values
- Invalid index record grooming
- Keep AppScale logs between restarts
- Backup & Recovery of application source code
- Fixed AppController self-lock issue
- Fixed issue with AppScale not coming up after reboot
- Default monit configuration changes (naming, cleanup, file checks, etc.)
- Fixed bug in removing applications from AppScale

AppScale version 2.4.0, released September, 2015
Highlights of features/bugs in this release:
- Upgraded to Ruby 1.9
- Upgraded Java support to SDK 1.8.1
- Fixed Datastore bug: fetching fewer references than the existing
- Fixed Datastore bug for values containing a reserved delimiter
- Fixed Datastore bug: failed transactions resulting in deleted index entries
- Allow multiple equality filters for single property and zz merge join queries
- FastStart for Google Compute Engine
- Added Shibboleth login capability
- Backbone of Hermes, AppScale's Messenger
- Centralized application log on the head node
- Reporting the ports used by an application
- Better AppScalefile argument parsing
- Better encoding handling for Java XML files
- Fixed Java bug in connecting to the Memcache service
- Fixed Java bug with cron entries not found
- Added BlobInfos implementation in Java Blobstore API
- Removed confirmation page after AppScale login
- Changed bootstrap to build from latest release by default
- Prevent Cassandra from filling up the disk with heap dumps
- Be more aggressive with log rotation
- Added email layer that enables debug mode

AppScale version 2.3.1, released April, 2015
Highlights of features/bugs in this release:
- Remove old dashboard data and add timestamps
- Release locks on soft deletes

AppScale version 2.3.0, released April, 2015
Highlights of features/bugs in this release:
- User management scripts
- Whitelisted Crypto libraries in python
- ZooKeeper autopurge
- Separated out data grooming service
- Crontab update fix
- Fixed functional tests for soap server
- Fixed ZK blacklist check
- Fixed handling application specified environment variables
- Validate crontab lines before inserting
- Removed secret from command line args for application servers
- Use hash of secret for RabbitMQ
- AppController is now a system service
- Improve error messages in EC2
- Removed duplicate entities when doing queries on list properties
- Better handling on AppScale versions
- Removed unused scripts

AppScale version 2.2.0, released Febuary, 2015
Highlights of features/bugs in this release:
- Updated NDB to latest version
- GHOST patch
- No more uncommitted reads outside or inside a transaction
- PHP version fix
- Fixed unit test
- Initial support for Search API (experimental)
- Fix for "+" sign in EC2 key for multinode deployments
- Updated fast start
- Removed unused scripts and deadcode
- Removed apache from build
- Removed ntp call in cronjob
- Updated logic for reconnection to ZK
- Fixed FD leak for ZK
- Native backup and restore for AppScale
- Catching exception from Cassandra causing groomer/garbage collection issues
- Fixed issue when binding to private IP
- Made unit tests run faster
- Added support for all ancestor paths on composite indexes

AppScale version 2.1.0, released December, 2014
Highlights of features/bugs in this release:
- Updated bootstrap script
- No longer overwriting SSL cert and private key on "appscale down"
- NDB patch
- Remove duplicates of host names found in ZooKeeper
- Better logging in AppController
- Use default HAProxy health checking
- Allowing connections to self signed certificates for remote API
- Email fix to allow for multiple attachments (python)
- Removed API checker
- Script to enable datastore viewer by IP
- White listed future builtins
- No longer storing temporary queues
- Reloading taskqueue queues on redeploy
- Increased concurrency of taskqueue celery workers to 10
- Handle new queues when queue definitions change
- Provide an empty cursor if requested
- Force HTTPS when pointing to port 443 (python)
- Dashboard navigation fixes
- Ancestors for composites now always point to the root entity
- Fixed maximum number of groups in a XG transaction
- Catch exceptions for ZK errors in datastore server
- Throw correct exceptions on illegal XG operations (needed for objectify)
- Java has its datastore set to HighReplication
- Added OFair to celery workers to not prefetch tasks
- Increased the number of datastore servers to be a multiple of CPU cores
- Improved groomer to handle more errors and retry
- Disable ssl3 (POODLE)
- Have Groomer delete logs older than 7 days
- Clean up transaction journal in groomer
- More fault tolerance with datastore operations and better load balancing
- Fixed broken unit tests
- Log rotation for saving disk space
- Remove conflicting jars when uploading a Java application
- Make sure jars are copied in when a java application is re-uploaded
- Remote API support for java (custom jar to include in application)
- Retry logic for sending emails (python)
- Have a java error application when an upload goes bad
- Fast start script for easy single node deployment
- RabbitMQ fix for node clustering

AppScale version 2.0.0, released June, 2014
Highlights of features/bugs in this release:
- Better error messages for AppScale tools
- Moved to package installation of most installed software (faster build)
- Cleanup of instances on "appscale down"
- Ability to support more dynamic code layout of Java applications
- Upgraded Cassandra to 2.0.7
- Use ntp service for clock sync
- Have AppScale catch when there are not enough resources on start
- Using openjdk instead of Oracle's JVM
- Fixed issue where monit thought ZooKeeper was not running
- Made build process more resilient
- Dead code elimination
- Wheezy support
- Updated bootstrap script
- Batching for datastore queries (better memory management and stability)
- Support for list composite indexes
- Support for distinct queries
- Projection queries no longer fetch from the entity table (faster queries)
- ZigZag Merge Join optimization
- Taskqueue statistics now reports number of tasks pending in a queue
- Fixed critical bug where a failed slave node would come up as master
- Fixed critical bug with nginx failing to be reloaded
- Experimental OpenStack agent (auto scaling)

AppScale version 1.14.0, released on February 10, 2014
Bugs fixed in this release:
- Removed "s~" character from app IDs (not used in AppScale) (#1428)
- Updated message for MapReduce library (#1427)
- Updated Nginx to latest version 1.5.10 (#1412 and #1411)
- Releasing orphan locks in ZooKeeper (#1410)
- Fixed AppController unit test (#1409)
- Ability to use datastore_admin builtins in AppScale (#1408)
- Cleaned up java AppServer log (#1407)
- Removed unused scheduler in Java TasQueue due to Null Pointer (#1406)
- Increased Cassandra timeout for bigger results from Datastore API (#1405)
- Async logging and more efficient log updates (#1397 and #1403)
- Allow the ability to update the flower password from the AppScalefile (#871 and $864)
- Increased the number of cycles before killing a process (#868)
- Ability to use remote_api in AppScale (#867)
- Fixed unit test for ZooKeeper retry (#866)
- Fixed Java logging (#865)
- Ability to set max memory for application in AppScalefile (#862)
- Exposing monit dashboard in cloud admin console (#861)
- Fix for null values in composite queries (#860)
- Fixing Channel API because AppScale secret changes on redeploy (#859)
- Fixing ZigZag Merge Join queries (#858 and #850)
- Setting max limit to 10k for queries to match GAE (#857)
- Having Channel API return 200s instead of 503 timeouts (#856)
- Cleaning up left over locks (#855)
- Using Nginx port instead of server port where appropriate (#855)
- Fix for multiple filters on a property in composite queries (#851)
- Fix parsing app.yaml for regex for nginx (#849)
- Setting PYTHON_LIB for app.yaml (#848)
- Always rewrite nginx config file on app upload (#847)
- Changed kind delimiter from "!" to 'x/01' (#844)
- Increased cutoff time for all paths to 10 minutes (#843)
- Channel API was missing named arg (#842)
- Notify failed transaction on datastore connection issues (#838)

AppScale version 1.13.0, released on December 19, 2013.

Bugs fixed in this release:
- Composite queries in the Java AppServer should make use of composite indexes
- Users should be able to specify an elastic IP address in their AppScalefile for their head node in Amazon EC2
- Users should be able to specify a static IP address in their AppScalefile for their head node in Google Compute Engine
- Users should be able to deploy applications written with the Go 1 runtime.
- Upgrade ZooKeeper to the latest version.
- pycassa client should be invoked with a list to all database peers, instead of only one peer
- Task Queue API calls should not fail on Eucalyptus, on any number of nodes
- Update Google Compute Engine interface to use v1beta16, since v1beta15 is now deprecated.
- AppController shouldn't write an empty file for API status if it didn't update API status info
- Apps should be able to detect if they are being accessed via HTTP or HTTPS
- Users should be able to use a command-line tool to get and set AppController-specific parameters
- scp flag should rsync over AppServer_Java directory
- AppScalefile template for cloud should include info about static IPs
- Python AppServer should not send empty logs to the AppDashboard
- APIs listed on status page should be displayed in a consistent order
- Fix JPEG decoder error.
- Fix broken IaaS Manager unit tests
- Application redirects from SSL port to non-SSL port, even if the path is set to always encrypted.
- RabbitMQ should not erase all persisted task queue data on disk when starting up
- Add button on Dashboard to go to #appscale
- Using latest version of Celery crashes workers on startup, so use stable version instead
- SOAP timeouts between AppControllers should not cause AppScale to crash or become unresponsive
- AppScale Tools should warn the user if they attempt to run AppScale on a machine with not enough memory for Cassandra
- Use the newest version of boto instead of locking to version 2.6
- 'appscale status' and AppDashboard should not show information about machines from previous runs when elastic IPs are used
- AppScale should start up fine even if the zookeeper user does not initially own the directory where ZooKeeper stores data
- AppController should not crash if APIChecker data is malformed
- Move button to download logs to Log Viewer page, instead of being on Status page
- Correctly calculate CPU and memory usage from top
- AppController should not ping the APIChecker every 20 seconds, but instead once every 5 minutes
- Verify that the loss of a single database peer in an AppScale deployment does not crash AppScale, while other peers still live
- AppScale should not crash intermittently on EC2 with "Cannot find my nodes in list of nodes" message
- Downscaling should kill all old dev_appserver processes, and should politely kill dev_appservers before SIGKILL'ing them
- clear_datastore flag should cause data to be consistently erased on startup on VirtualBox deployments
- AppScale should start up fine in Amazon EC2 if persistent disks get attached at somewhere other than the requested location
- The autoscaler should not add AppServers to machines that have more than 90% CPU used
- Users should be able to tell how many requests per second hit their app when data is received for it
- Visiting the / URL on the Dashboard should go to the status page, not the landing page.
- Users should be able to send commands to be executed on AppScale VMs that include redirection characters
- If a machine is rebooted in an AppScale deployment and its IP address changes, it should not crash when restarting the AppController

AppScale version 1.12.0 was released on September 30, 2013.

Bugs fixed in this release:
- Upgraded GCE image to Ubuntu Precise
- Upgraded KVM image to Ubuntu Precise
- Upgraded VirtualBox image to Ubuntu Precise
- Upgraded Eucalyptus image to Ubuntu Precise
- Upgraded EC2 image to Ubuntu Precise
- Now using monit instead of god to monitor processes, and killing App Engine apps that take up too much memory
- Now using monit to revive nginx if it fails
- Now using monit to revive nginx if it fails
- Now using monit to revive rabbitmq if it fails
- Now using monit to revive ZooKeeper if it fails
- Now using monit to revive haproxy if it fails
- Now using monit to revive ejabberd if it fails
- Now using monit to revive Cassandra if it fails
- Improved Blobstore API fidelity for Python 2.7 apps
- XMPP now works correctly on relocated apps
- Load is rebalanced after scaling events
- AppScale can now run in other AWS regions
- No longer displaying oauth2client warnings on GCE
- Not failing GCE image creation due to modprobe not being supported
- Logging levels off by one in AppDashboard for Python apps
- Fixed broken ImageLoadTest on Python 2.7 Hawkeye
- Java AppServer now reads nginx port from filesystem, instead of as an arg
- Protect access to flower
- Nginx shouldn't expose Blobstore port to users in their apps
- Prevent deprecated Python 2.5 apps from being uploaded
- Allowing users to set keys longer than 250 characters in Java Memcache
- AppServer ports are now automatically reused
- Upgraded Cassandra to 2.0.1
- Nginx and haproxy ports are now automatically reused
- If there is a version mismatch between tools and main, tools now report each version to user and what they can do to fix it
- AppController now gets IP correctly on non-English locales
- Users can now bootstrap AppScale VMs with a list of commands in their AppScalefile
- kill_all_instances_for_app now works every time, instead of just the first time
- Celery works on rebooted VMs in one node deployments
- Cleaned up AppDashboard interface to AppController
- Fixed uploading apps via the AppDashboard
- Removed link to Admin Console in AppDashboard
- Uploading apps via the Dashboard now shows their information there
- AppDashboard now allows users to upload zip files
- Remove app reliably clears apps off the Dashboard
- InfrastructureManager no longer prints users' AWS secret key
- Removing apps then uploading them no longer results in 500s
- Dashboard now shows updated ports for apps after relocation
- Uploading apps via the Dashboard that were removed no longer results in 500s
- Command-line interface now allows users to upload zipped apps
- Upgraded pycassa to 1.9.1

AppScale version 1.11.0 was released on September 30, 2013.

Bugs fixed in this release:
- Updated license to Apache 2.0
- Use devappserver2 for multithreaded Python 2.7 apps
- (Experimental) AppScale should run on Ubuntu Precise
- Enable users to scale up from a single node
- Allow users to dynamically change the port their app runs on
- Support PHP App Engine apps
- Do not let users start AppScale if there are terminated instances with the same keyname
- Providing AppScale with invalid credentials on EC2 or Eucalyptus should fail with a more descriptive error message
- Validate that XMPP, Cron, and Task Queue works on relocated apps
- Fix bulk loader to use new ports for AppDashboard
- Upload app page should include text explaining how to upload apps by default
- 'appscale tail' should tail files across AppScale deployments
- Relocating apps should not cause upload-app to return the wrong port on multinode deployments
- Disable app relocation to firewalled ports
- AppController should persist AppServer info to ZooKeeper
- Relocating apps should result in the correct port being propogated to the AppDashboard
- Removing apps shouldn't cause 502s from nginx
- Give users a better error message if the keyname or group is in use
- 'appscale down' then 'appscale up' should automatically restart apps on the same ports
- Make sure logs get sent to the AppDashboard on Python 2.7 apps
- Add relocate info to AppScale tools usage
- Don't use dashes in keyname and group, since GCE doesn't allow it
- AppDB should reestablish ZooKeeper connections when they are broken
- Don't use appscale.cs.ucsb.edu in blobstore connection string
- Nginx should serve static files with spaces correctly
- Fix Python MapReduce API
- Fidelity fixes for Java Memcache API
- Java App Engine should indicate that its server mode is 'Production'
- Support for Zigzag queries
- Remove TimesTen build files
- Upgrade API checker to Python 2.7
- Use stack size of 226k with Cassandra
- Use updated port on API checker when performing urlfetch on AppDashboard
- Cloud admins should be able to view the App Console for all apps, not just apps they own
- 'appscale down' or 'appscale clean' should remove local AppController state
- Not displaying unknown role on status page
- 'appscale --version' should print version number instead of usage
- lxml should be importable for Python 2.7 apps with new devappserver
- Sleep after VMs have been started in GCE to allow for SSH key injection
- 'appscale up' and 'appscale down' check the same files to see if AppScale is running
- Require Python 2.7 when using GCE, since the Google library requires it
- Include gflags in build script

AppScale version 1.10.0 was released on August 28, 2013.

Bugs fixed in this release:
- Automatically persisting data across AppScale runs in EC2/Eucalyptus if EBS volumes are provided
- Automatically persisting data across AppScale runs in GCE if PD volumes are provided
- Automatically persisting data across AppScale runs in VirtualBox
- AppScale now restarts if machines were halted or powered off
- Allowing users to specify what availability zone instances are spawned in
- Serving static files via nginx for Java App Engine apps
- Simplified autoscaling system
- Passing over more detailed information about why the AppController crashes when a non-recoverable error ocurrs
- Improved Java Datastore API fidelity
- createLogoutURL in Java Users API now redirects users back to the specified URL
- 'appscale clean' no longer sometimes requires 'force: True' on subsequent 'appscale up's
- Added support for CAS in Python Memcache API
- Politely shutting down database to avoid data loss / corruption
- Enabling tools to take in signed OAuth2 credentials, so that web apps can use the AppScale Tools with GCE
- Added ability to reset user passwords via the Dashboard
- Verbose: True in AppScalefile now produces more logging in the AppController automatically
- Enabling users to import Remote API Shell in their apps
- Fixed memory leak where apichecker and appscaledashboard apps would endlessly collect logs
- Fixed Task Queue API to not enqueue tasks with the same name as previously enqueued tasks
- Tools no longer crash if they can't set the locale
- Adding keyname and group in AppScalefile is no longer whitespace sensitive
- AppController no longer crashes if cron.yaml is invalid
- Updated AppScalefile templates to include new flags
- No longer serving tar.gz'ed apps in the apps directory, since it could be accidentally served to users in some cases
- Storing apps in /opt/appscale/apps, so that we persist them across AppScale deployments
- Specifying toPort and fromPort when authorizing ICMP traffic (to work on Eucalyptus 3.3)
- 'appscale logs' now works if the AppController has failed
- Using IP addresses in Eucalyptus deployments instead of DNS names.
- AppDashboard no longer crashes if the Console page is viewed while an app is loading
- Added link to App Engine Admin Console on AppDashboard
- No longer reporting negative requests per second in AppDashboard for apps
- Stripping leading and trailing whitespace on username when users log into AppDashboard
- AppDashboard no longer infinitely retries when talking to AppController
- Create random keyname and group 
- Link to header in AppDashboard now works again
- AppController retries requests to AppDashboard that fail b/c of transient errors
- Updated GCE support to use v1beta15 instead of the now unsupported v1beta14
- Telling users that appscale clean deletes their data
- Providing ability to remove nameservers from /etc/resolv.conf to potentially speed up IP resolution
- appscale deploy yields a more useful error message if users give it something that isn't a directory or .tar.gz file
- Kind stats in AppDashboard have hover text, to make it more readable when many entities are present
- AppScale no longer allows apichecker or appscaledashboard to be used as appids by users
- Removed m2crypto from OS X build script
- Installing dig on VMs
- Terminating GCE instances in parallel, instead of in serial

AppScale version 1.9.0 was released on July 22, 2013.

Bugs fixed in this release:
- Autoscaling within machines for Python and Java App Engine apps
- Autoscaling to new machines for Python and Java App Engine apps
- Downscaling for Python and Java App Engine apps
- Updated Cassandra to 1.2.5
- Java App Engine fixes to improve API fidelity with App Engine
- AppController shouldn't crash if cron.yaml specifies no jobs
- AppController logs shouldn't print 'warning: peer certificate' messages
- Re-uploading same application shouldn't result in duplicate cron jobs
- Blobstore should parse string entities correctly
- AppController shouldn't crash if no queues are specified in queue.yaml
- InfrastructureManager should report when requests fail
- Add memcache stats to AppDashboard
- Add flower link on AppDashboard
- TaskQueue should not report status as failed when it is working fine
- AppDashboard should show new nodes on status page when upscaling occurs
- Increased number of concurrent requests to VMs
- Handling SSH keys more intelligently in GCE
- Handle Unavailable exceptions better with pycassa
- Installing wget before it is used in build script
- AppController shouldn't crash if other AppControllers are down
- AppDashboard should list instance information for AppServers
- XMPP receiver should reconnect to ejabberd if it loses its connection
- Ensure all AppController logging goes to AppDashboard
- Parsing queue.xml file for Java App Engine apps
- Added ability to generate kind statistics on demand in AppDashboard
- IP addresses no longer disappear when window resize occurs in AppDashboard
- AppController shouldn't crash if it can't find app.yaml or appengine-web.xml
- Added Java 7 into AppScale image

AppScale version 1.8.0 was released on June 17, 2013.

Bugs fixed in this release:
- Google Compute Engine support
- Upgrade Python App Server to support 1.8.0 APIs
- Upgrade Java App Server to 1.8.0
- Improved performance of ZooKeeper within AppScale
- Displaying Kind stats for each app in Dashboard
- Displaying requests per second for each app in Dashboard
- Remove MySQL from AppScale images
- AppScale Tools now writes an informative crash log if an uncaught exception is thrown
- Fixed Channel API on 1.8.0 Python App Server
- Apps should be able to specify queues without a rate specified.
- Use private IPs when storing ZooKeeper locations
- Fix install script so that 2nd install attempt will pass
- Pushing tasks to queues that don't exist no longer kill the Task Queue Server
- Enabling users to specify queue names with dashes in them
- Java App Server properly checks if user cookies are valid before telling user they are / are not logged in
- Task queue tasks can be run as admin in Java App Server
- Fixed regression where app list refresh was broken, and JSON was being printed to status page
- Added clarification on how to use force flag in AppScalefile
- Fixed broken continue URLs in Java App Server
- Better exception handling for ZooKeeper client in AppDB
- Moved ZooKeeper data to /opt/appscale/zookeeper
- Moved Cassandra data to /opt/appscale/cassandra
- Fixed regression where groomer was broken
- Left aligning logs in Dashboard
- Properly rendering Dashboard icons
- AppController now starts successfully even if other failed AppController processes are present
- Dashboard now handles compressed file uploads with single directory better
- Removed datastore_writes from API Checker and Dashboard
- Updated cloud AppScalefile template to tell users how to use it with GCE
- Printing crash log info more intelligently

AppScale version 1.7.0 was released on May 13, 2013.

Bugs fixed in this release:
- System-level logging for admin user
- Fixed 'none' in database name, replication, and monitoring URL
- Speed up performance of AppDashboard
- Implement log levels for the AppController
- Prime AppDashboard before sending users to it
- Extend timeout from 30 to 60 minutes for images to boot in Eucalyptus.
- Add AppStats support for Python 2.7 App Engine apps.
- Refactor AppController's loging to not have everything at the DEBUG level.
- Namespacing now handled correctly by the datastore server.
- Uploading large apps from web UI now succeeds
- Updating login cookie at AppDashboard when apps are uploaded or deleted.
- Static secure files no longer lead to redirect loops.
- Channel API support for Java App Engine apps.
- Remove runtime checks for invalid Java classes from SDK.
- Mail API support for Java App Engine apps.
- AppDashboard now stores API status in Datastore
- AppController dying no longer kills the AppDashboard
- Resolved ndb deadlock issues with AppDashboard
- Fixed Cron API on Eucalyptus
- Altered build script to fail if any component fails to install
- Fixed 'appscale down' / terminate-instances when running on an AppScale image
- Apps can now be updated without first having to remove them
- Using Kazoo for ZooKeeper interactions, fixing NFS stale file handle errors
- Added EC2_SECRET_KEY and EC2_ACCESS_KEY args for AppScalefile
- Removed unused JavaScript files in AppDashboard
- 'verbose: False' now omits verbose flag in 'appscale up'
- Bulk loader now generates kind statistics
- Added bulk loader support for Java App Engine apps
- Java App Engine apps now support multithreading
- AppDashboard now uses graphical representations of numeric data
- Fixed XMPPReceiver, which broke with new AppDashboard implementation
- Cleaning up soft deleted items periodically
- ndb no longer sees memcache entries as corrupt
- Removed update SDK message from Java App Server
- Fixed bug where XMPP would fail for Java App Engine apps with capital letters in appid
- appscale.deploy now returns the host and port of the app that was deployed
- Enable AES support for apps that use pycrypto
- Properly adding lxml to dev_appserver's import path
- Removed AppDashboard's font references in CSS to files that don't exist
- Allow users to upload tar files with the app in a top-level directory
- When doing an 'appscale clean', we now remove local state in ~/.appscale
- Fixed starting AppScale on EC2 with a one node deployment, with --ips_layout
- Using retries when performing Kazoo operations with ZooKeeper
- appscale-upload-app now uses unique tempdir locations when uploading apps
- Uploading apps via web UI no longer throws 404s
- Large apps no longer timeout nginx when uploaded via the AppDashboard
- Using glob when importing Python libraries in dev_appserver
- Refreshing application names consistently in AppDashboard
- Caching AppController status at login node, resolving SOAP timeouts in Dashboard
- Improved stability for AppDashboard in Euca deployments

AppScale version 1.6.9 was released on April 5, 2013.

Bugs fixed in this release:
- Cross group (XG) transactions for Python 2.5, Java, and Python 2.7 App Engine apps
- Java XMPP support
- Spot instance support when running over Amazon EC2
- Automatic price estimation algorithm for spot instance support
- Robustness fixes for ZooKeeper interface
- Transactional task support for Python App Engine apps
- Memcache increment / decrement in Java App Engine
- Fixes for Python App Engine Channel API
- Have god monitor HAProxy
- NTP sync for VMs in AppScale deployments
- 'appscale deploy' now works for relative app paths
- Deferred task support
- Increased validation on AppScalefile
- Rebranded continue URL page
- "appscale clean" command, to terminate AppScale on all nodes
- Fixed "submit issue" button on dashboard
- Added support for environment variables in app.yaml files
- Starting AppMonitoring service in EC2 correctly

AppScale version 1.6.8 was released on March 13, 2013.

Bugs fixed in this release:
- Ordered ancestor queries
- Complex query fix for indexes with same name
- Obscured EC2 credentials in AppController log
- Turn off firewall when we turn off AppScale
- New TaskQueue client for Java
- Fixed XMPP API when running on EC2, Eucalyptus
- Fixed XMPP API for Python 2.7 App Engine
- Fixed errors using AppScale on complex deployments
- Rebranded AppLoadBalancer
- Fixes for distributed Task Queue system
- Restarting RabbitMQ if it fails to start up
- Restarting AppControllers on remote nodes if they fail to start
- Added pycrypto and lxml for use by App Engine apps
- Fixed timeout for RabbitMQ starting up to not be infinite
- Added deferred Task Queue support
- Replaced spymemcached with xmemcached for Java App Engine

AppScale version 1.6.7 was released on February 13, 2013.

Bugs fixed in this release:
- Terminate script kills Erlang processes
- Fixed HTTP -> HTTPS redirection issues
- Add support for JDO and JPA
- Properly encoding memcache keys with spaces in it
- Fixed Java memcache inconsistencies in a distributed environment
- Changed nginx config file construction to support regexes

AppScale version 1.6.6 was released on January 21, 2013.

Bugs fixed in this release:
- Support Python 2.7 App Engine apps
- Fixed ACID semantics for transactions
- Add nodes dynamically to a running AppScale deployment
- SSL support for App Engine apps
- Upgrade nginx to latest stable version and include HTTP chunking support
- AppController revives properly after being killed
- Cookies do not get created in certain deployment types
- Cookies do not get deleted if > 1 eth device present
- Turn on HRD flag in Python AppServer
- Added AppScalefile support, making it easier to run AppScale
- Add in 'appscale logs' command to automatically collect log files
- Add in 'appscale tail' command to automatically tail log files
- Add in 'appscale down' to alias to 'appscale destroy'
- Add in 'appscale ssh' to ssh to AppScale VMs
- Added Watchfile for continuous testing support
- Remove binary from beginning of locations.yaml file
- Start unrelated API services in parallel to improve startup time
- Make tools print version number on startup
- Fix bulkloader to support batching
- Python and Java API fidelity testing apps (Hawkeye)
- Obscuring EC2 credentials after a regression exposed them

AppScale version 1.6.5 was released December 19, 2012.

Bugs fixed in this release:

- Update Java AppServer to 1.7.3
- Update Python AppServer to 1.7.3
- Make bulkloader work with AppScale
- AppScale fails to start if eth0's IP is not externally accessible
- NeptuneManager no longer starts up correctly
- AppController should validate AppScale version
- Firewall needs to be on
- Hypertable does not start when db_master != node1

AppScale version 1.6.4 was released November 21, 2012.

Bugs fixed in this release:

- Update Java AppServer to 1.7.2.1
- Python AppServer leaks memory
- InfrastructureManager is not using given user's credentials
- AppController crashes if it can't determine its own IP address
- Failure to delete app data from ZooKeeper prevents app removal 
- Uploading apps via web UI doesn't work 
- Removed sisyphus from code base 
- New builds fail due to Rake error 
- AppScale lite - selectively build one database into AppScale instead of all of them

AppScale version 1.6.3 was released October 30, 2012.

Bugs fixed in this release:
- Autoscaler scales down when --appengine is used (https://github.com/AppScale/appscale/issues/43)
- AppController crashes in response to failed ZooKeeper operations (https://github.com/AppScale/appscale/issues/47)
- Python applications should start even if rabbitmq fails to start (https://github.com/AppScale/appscale/issues/62)
- Bad applications crash appscale (https://github.com/AppScale/appscale/issues/57)
and updated versions of Cassandra and HBase used.

AppScale version 1.5 was released July 28, 2011. Features include:

	Support for the bulkloader, enabling persistence for your data
	Upgraded Java and Python AppServers to GAE 1.4.3
	Support for Go App Engine apps (SDK version 1.5.0), including support for apps that use multiple processes
	Fault tolerance for almost all services (processes monitored and revived by god)
	Faster startup and termination of AppScale, especially over larger numbers of nodes
	Tools and image now verify that all instances used have AppScale installed
	EC2 and Eucalyptus credentials are now obscured when they are printed to logs
	Channel API for Python (multiple receivers can also be used) - implemented via Strophe.js
	Blobstore and Files API for Python
	XMPP API for Python - implemented via ejabberd
	Hybrid cloud support - run AppScale over multiple clouds in a single deployment (e.g., Eucalyptus and EC2, EC2 East Coast and EC2 West Coast)
	Neptune language support
	Table caching for MySQL, HBase, Hypertable to improve performance
	Updated interface for Amazon SimpleDB
	Upgraded Cassandra version used to 0.7.6-2
	Upgraded HBase version used to 0.89
	Upgraded Hadoop version used to 0.20.2
	Upgraded Hypertable version used to 0.9.43
	Namespacing support
	Added Loki, a fault tolerance tester along the lines of Netflix's Chaos Monkey
	User authorization system for MapReduce, EC2, and Neptune APIs
	Ability to remove transaction overhead via namespaces
	Various other bug fixes

AppScale version 1.4 was released September 7, 2010.

Included in this release (over version 1.3) is:

    Transaction support for all databases
    Placement support - administrators can now specify which components should go where
    Package installation via apt-get for Ubuntu Jaunty, Karmic, and Lucid
    Addition of support for Scalaris
    Upgrade to Python App Engine 1.3.2 compatibility
    Introduction of Java App Engine 1.3.5 compatibility
    Added support for Images API and Cron API for Python Google App Engine
    Added support for Memcache API, Images API, Task Queue API, and Cron API for Java Google App Engine
    Single node deployments for all databases (previously only Cassandra, Voldemort, MongoDB, and MemcacheDB were supported)
    PBServer is now multi-process and load balanced to improve throughput
    Secure login page for authentication via https
    Amazon AWS EC2 public image available (ami-044fa56d) 

Supported Databases:

    HBase - Version 0.20.3
    Hypertable - Version 0.9.2.7
    MySQL Cluster - Version 5.1.30
    Cassandra - Version 0.6.1
    Voldemort - Version 0.80
    MongoDB - Version 1.2.2-1ubuntu1
    MemcacheDB - Version 1.2.0-6
    Scalaris - Version 0.2.3-1 

Restrictions:

    64 bit systems only (host and virtual machines)
    Specific versions supported (we support only these versions)
        Ubuntu Jaunty, Karmic, and Lucid only
            x86_64 image kernel/modules/ramdisk 
        Xen v3.3.3
        Eucalyptus v1.6 (KVM and Xen virtualization supported)
        Python App Engine 1.3.2 and Java App Engine 1.3.5
        HBase, Hypertable, MySQL, Cassandra, Voldemort, MongoDB, MemcacheDB, Scalaris 

Roadmap

    Persistence of instance data via EBS in Eucalyptus and EC2
    Automatic scaling of AppScale deployment in Eucalyptus/EC2
    Additional robustness and bug fixes 

Prior Releases

Release 1.3, released December 13, 2009.

Included in this release (over version 1.3 Beta) is:

    Critical bug fixes for Cassandra, MongoDB, and MySQL
    Amazon AWS EC2 public image available (ami-14799b7d) 

Restrictions:

    64 bit systems only (host and virtual machines)
    Specific versions supported (we support only these versions)
        Ubuntu Jaunty 9.04 distribution supported only
            x86_64 image kernel/modules/ramdisk
            for virtualized instances:
                linux-ubuntu-modules-2.6.24-23-xen
                linux-restricted-modules-2.6.24-23-xen
                linux-image-2.6.24-23-xen 
        Xen v3.3.3
        Eucalyptus v1.5.2 (KVM and Xen virtualization supported)
        Python App Engine 1.2.7
        HBase, Hypertable, MySQL, Cassandra, Voldemort, MongoDB, MemcacheDB (versions specified in the scratch install instructions) 

Release 1.3, Beta, released December 7, 2009:

    Upgrade to Python App Engine 1.2.7 compatibility
    Introduction of Java App Engine 1.2.5 compatibility
    Addition of support for MongoDB and MemcacheDB
    AppServers are now multi-process and PBServer is now multi-threaded, greatly improving performance
    True memcache(d) support for Python App Engine apps
    MySQL bug fix allowing runs over any number of nodes greater than one
    Amazon AWS EC2 public image available (ami-5e05e737)
    New single node deployments for MongoDB and MemcacheDB
    Ability to delete applications from a running deployment
    Python2.6 for everything except Google App Engine (Python2.5)
    Complete rewrite of AppLoadBalancer
    True support for static files via nginx (Python App Engine apps only)
    Numerous security fixes relating to cookies and administrator status on apps
    Additional robustness and bug fixes 

Restrictions:

    64 bit systems only (host and virtual machines)
    Specific versions supported (we support only these versions)
        Ubuntu Jaunty 9.04 distribution supported only
            x86_64 image kernel/modules/ramdisk
            for virtualized instances:
                linux-ubuntu-modules-2.6.24-23-xen
                linux-restricted-modules-2.6.24-23-xen
                linux-image-2.6.24-23-xen 
        Xen v3.3.3
        Eucalyptus v1.5.2 (KVM and Xen virtualization supported)
        Python App Engine 1.2.7
        HBase, Hypertable, MySQL, Cassandra, Voldemort, MongoDB, MemcacheDB (versions specified in the scratch install instructions) 

Release 1.2, released September 7, 2009:

    Upgrade to Python App Engine 1.2.3 compatibility
    Addition of support for Voldemort and Cassandra
    MySQL bug fix allowing for parallel API nodes
    Eucalyptus 1.5.2 support (no previous versions of Eucalyptus are supported)
    Amazon AWS EC2 public image available (ami-7136d618)
    Support for running Ubuntu Jaunty systems
    Single node deployments (Cassandra or Voldemort only)
    Ability to delete applications from a running deployment
    Python2.6 for everything except GAE (Python2.5)
    Replication is configurable
    Ubuntu 9.04 support and wiki directions
    Wiki directions of sole-KVM use
    Additional robustness and bug fixes 

Release 1.1, released June 17, 2009:

    Upgrade to Python App Engine 1.2.2 compatibility
    Addition of support for MySQL
    Eucalyptus 1.5 support (no previous versions of Eucalyptus are supported)
    Amazon AWS EC2 support and instructions
    Instructions for building an AppScale image from scratch
    Support for running on non-virtualized Ubuntu Hardy (8.04) systems
    iptables-based firewall for improved security
    Additional robustness and bug fixes 

    Restrictions:

        64 bit systems only (host and virtual machines)
            32 bit systems for non-virtualized installation 
        Specific versions supported (we support only these versions)
            Ubuntu Hardy Heron 8.04 distribution supported only
                x86_64 image kernel/modules/ramdisk
                for virtualized instances:
                    linux-ubuntu-modules-2.6.24-23-xen
                    linux-restricted-modules-2.6.24-23-xen
                    linux-image-2.6.24-23-xen 
            Xen v3.3.1
            Eucalyptus v1.5 (KVM and Xen virtualization supported)
            Python App Engine 1.2.2
            HBase, Hypertable, MySQL (versions specified in the scratch install instructions) 

Release 1.0.2:

    Supported Databases: HBase, Hypertable
    Cluster types supported: Xen, Eucalyptus
    Bug fixes (1.0, 1.0.1) 

Release 1.0:

    Xen-only, image-based AppScale deployment released
    Supported Database: HBase