Fwd: 1.6.4 Prerelease Now Available

[Marzia Niccolai]

Hi,

The 1.6.4 prerelease SDKs are now available to download for Python and Java here: http://code.google.com/p/googleappengine/downloads/list

Please see release notes below.

-Marzia

Python Version 1.6.4
===============================
- Datastore statistics now show the amount of storage used by application
 indexes.
- Billed applications which have specified additional logs storage over 1 GB
 are now being charged for that storage at $0.24/GB/month.
- We have released an experimental utility for backing up your application's
 blobs at the same time you migrate your datastore data. You can opt in to
 blob migration in the Admin Console when you start your migration.
- Using the Blobstore API's serve_blob method, your application can serve
 objects hosted on Google Storage for Developers.
- We have updated the experimental Backup/Restore functionality to include
 the option to backup and restore to Google Storage for Developers.
- The NDB datastore API is now generally available.
- In the Python 2.7 runtime, Background threads are available as an
 experimental release when using App Engine backends.
- In an application's Admin Console, there is a Memcache viewer available
 that lists Memcache stats and has the ability to view Memcache content
 based on key.
- In the Capabilities API stub in the SDK, you can now enable or disable
 a capability using SetPackagedEnabled.
- The Windows installer now prompts to install Python 2.7 instead of
 Python 2.5.
- The Testbed API now supports the Capabilities API.
- GQL queries in the Admin Console no longer throw an error when a trailing
 semi-colon is included.
- The Datastore API now includes a NonTransactional decorator to ensure that
 a function is run outside of a transaction. Existing transactions are paused
 while the function is executing.
- The Datastore Admin now shows entities from every namespace.
   http://code.google.com/p/googleappengine/issues/detail?id=3962
- Fixed an issue with _strptime when threadsafe was specified.
   http://code.google.com/p/googleappengine/issues/detail?id=6489
- Fixed an issue where DatastoreFileStub.__del__ fails on tempfile.msktemp.
   http://code.google.com/p/googleappengine/issues/detail?id=6749
- WebOb 1.1.1 is now included in the SDK, and used by default there when
 Python 2.7 is specified.
   http://code.google.com/p/googleappengine/issues/detail?id=7014
- Fixed an issue where the index.yaml file was cleared if your skip_files entry
 differs from the default skip_files list.
   http://code.google.com/p/googleappengine/issues/detail?id=7031

Java Version 1.6.4
=============
- Datastore statistics now show the amount of storage used by application
 indexes.
- Billed applications which have specified additional logs storage over 1 GB
 are now being charged for that storage at $0.24/GB/month.
- We have released an experimental utility for backing up your application's
 blobs at the same time you migrate your datastore data. You can opt in to
 blob migration in the Admin Console when you start your migration.
- Using the blobstoreService.serve method, your application can serve objects
 hosted on Google Storage for Developers.
- We have updated the experimental Backup/Restore functionality to include
 the option to backup and restore to Google Storage for Developers.
- Version 2.0 of the App Engine DataNucleus plugin is now available as an
 experimental release. This plugin contains support for JDO 3 and JPA 2 and
 over 40 bug fixes. For a full list see:
   http://code.google.com/p/datanucleus-appengine/source/browse/branches/2_0_0/dist/RELEASE_NOTES.ORM
- Threads are available as an experimental feature via
 com.google.appengine.api.ThreadManager.
- In an application's Admin Console, there is a Memcache viewer available
 that lists Memcache stats and has the ability to view Memcache content
 based on key.
- The method ApiProxy.getRemainingMillis() returns the amount of milliseconds
 remaining before a request hits the request deadline.
- New datastore query related callbacks - PreGet(), PreQuery(), and PostLoad()
 are now available.
- Omitting the <threadsafe> directive from appengine-web.xml now results in
 a warning. Starting with the 1.6.5 release, omitting this directive will
 result in an error.
- You can set an RPC deadline for the Task Queue API's leaseTasks method.
- GQL queries in the Admin Console no longer throw an error when a trailing
 semi-colon is included.
- Fixed an issue where <async-session-persistence> did not work.
   http://code.google.com/p/googleappengine/issues/detail?id=5774
- The Datastore Admin now shows entities from every namespace.
   http://code.google.com/p/googleappengine/issues/detail?id=3962
- Fixed an issue in the Java Development Server where Task Queue retry
 parameters were ignored.
   http://code.google.com/p/googleappengine/issues/detail?id=5200

[Tom Willis]

"- WebOb 1.1.1 is now included in the SDK, and used by default there when
 Python 2.7 is specified.
   http://code.google.com/p/googleappengine/issues/detail?id=7014"

/happy dance

[Jeff Schnitzer]

On Mon, Mar 19, 2012 at 8:10 PM, Marzia Niccolai
<marce+a...@google.com> wrote:
> - Omitting the <threadsafe> directive from appengine-web.xml now results in
>  a warning. Starting with the 1.6.5 release, omitting this directive will
>  result in an error.

This is a very good idea.

Jeff

[jon]

> - Fixed an issue where <async-session-persistence> did not work.
>    http://code.google.com/p/googleappengine/issues/detail?id=5774

Thanks for fixing this.

Can we implore Googlers to fix this too?
http://code.google.com/p/googleappengine/issues/detail?id=6236

[Ikai Lan (Google)]

I'm hoping you guys saw this:

- Billed applications which have specified additional logs storage over 1 GB
 are now being charged for that storage at $0.24/GB/month.

What's actually happening is that we are hiding logs, but once 1.6.4 is released, we will be deleting logs greater than 1gb. With 1.6.3 we released some settings for people to tweak Log Retention in Application Settings (http://googleappengine.blogspot.com/2012/02/app-engine-163-released.html).

This was a much requested feature, because prior to this, logs at each logging level were saved on a FIFO basis, forcing developers to do ugly things like regular log downloads.

--

Ikai Lan 
Developer Programs Engineer, Google App Engine

plus.ikailan.com

--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.
To post to this group, send email to google-a...@googlegroups.com.
To unsubscribe from this group, send email to google-appengi...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.

[Mos]

Please GAE-Team, stop developing features and take care of basic functionality and reliability of your platform.
New features doesn't make even sense if you can't stick to your basic non-functional promises (latency, reliability, scaling, performance).
For example:

http://groups.google.com/group/ggle-appengine-java/browse_thread/thread/83006c9480ee82b0?pli=1

and

http://code.google.com/p/googleappengine/issues/detail?id=7133

[Per]

- Billed applications which have specified additional logs storage over 1 GB
 are now being charged for that storage at $0.24/GB/month.

What's actually happening is that we are hiding logs, but once 1.6.4 is released, we will be deleting logs greater than 1gb. With 1.6.3 we released some settings for people to tweak Log Retention in Application Settings (http://googleappengine.blogspot.com/2012/02/app-engine-163-released.html).

Hi Ikai, definitely noticed. Excellent! Also, fairly modest pricing. Thanks!

 

[Robert Kluin]

Awesome. Tons of good features here.

[tempy]

I'm sorry but I think I have to back up Mos here. I'm as excited about
the new-shiny as anyone else, but no amount of new-shiny makes up for
the frayed nerves of annoyed users, instance-spawning chaos, requests
that never even reach my app, silence from google, a completely
oblivious status page, and so on. Please make the platform rock-solid
first, and then by all means, bring on the new-shiny.

On Mar 20, 1:10 am, Marzia Niccolai <marce+appeng...@google.com>
wrote:

> http://code.google.com/p/datanucleus-appengine/source/browse/branches...

[Jeff Schnitzer]

Appengine is a big team. There will always be some people working on
reliability while others work on features. This isn't an either/or
situation... and it probably doesn't help to dogpile everyone on
specific bugs.

Also: From observation, GAE has a deep queue of features that slowly
roll out to production. Chances are high that there *were* several
features that didn't make it into this release because the team was
busy chasing down problems. By past standards, 1.6.4 is pretty much a
yawner... at least, for the Java SDK.

Jeff

[Malte Jäger]

Just between us, when will Muggels get a permission to use:
HTTPS and FulltextSearch?

[PK]

I had to tweak my path to accommodate the new webob but then all our regression tests passed just fine.

Thanks for taking care to add functionality without regressing anything.

Still looking forward to having some of the aging bugs fixed and some of the long time coming roadmap items like "SSL for custom domains" released.

PK

[jon]

> Just between us, when will Muggels get a permission to use:
> HTTPS and FulltextSearch?

+1

[toonetown]

Yup - still waiting for resolution on deadlineexceeded exceptions...2+ weeks and counting. Even migrated to HRD (even though our app doesn't use data store at all)

Still nothing. And no way to contact anyone or find out what's going on.

The thing that pisses me off the most is if you look at the status dashboard, everything looks peachy - WHEN IT IS NOT!

[Brandon Wirtz]

how long is a standard request (Milliseconds)? How long is a the DEE
request?

What are your max pending latency and idle instance settings?

[Jeff Schnitzer]

Hmmm. Are you certain that the problem isn't on your end? When GAE
has a real issue there's a temptation to assume that every problem is
"the server's fault", even when it isn't. Your problem doesn't sound
like any of the other issues seen here, which typically have revolved
around instance restarts, failed instance startups (before user code
gets executed), and datastore latency.

Since you are seeing DeadlineExceededException, you must know what
call is causing the timeout. What service is blocking so long that
times out your app? Post a stacktrace.

Jeff