In my opinion,
Google, Facebook etc, they don't really care
On Feb 15, 4:29 pm, Max Harper wrote:
> Google fixed this issue last night. Thanks to the Google engineers who
> addressed the problem. I'm not sure this would have been addressed if
> we hadn't had a friend internal to Google who looke
well, i also agree to that.
I mean, my position is really clear. Im with you about the performance
just after the maintenance period (like right now im starting to reply
to all the complains for the last 4 hours today) - has been the worst
in the last 4 months.
was just being clear with a wider o
> 99.99% uptime? That's _expensive_ everywhere
I'd settle for even 90% right now. Today I'm nearing a 25% error rate
in the past 22 hours. That's pretty pathetic for a service. If it
was just one day I'd understand, but lately it's been nearly every day
where something has gone haywire (usuall
I also wanted to add this comment which is probably not helpful, but speaks
more to the root of the problem. :)
And that is this, "you can't get there from here". The fundamental problem
does not appear capable of being reduced beyond looking at a bunch of
Journeys. I'm going to summarize a coup
I am going to post two replies as I have a small idea that might help a
little bit, and then a further post on the whole problem in general.
First, Ed, what you have said is consistent and the problem of identifying
the "most recent Journey within a constrained time window" does appear to
imply
Hello Mr. wesley,
webapp.template.register_template_library('customfilters')
This is the only way we could import register_template_library,
customfilters.py file on the root level of the app, access from other paths
is not available.
We used the default package django 0.96 didnt try other vers
Taking a decision of leaving is actually a hard call. There are many
things involved ( + Business decisions)
I guess everything falls at what you expect from your service to end
users: 99.99% uptime? That's _expensive_ everywhere (or eats everyones
life at the sysadmin level resetting load balance
Hi Justin,
If you change your queue's bucket size to be 1 it will give you the
semantics you are after.
On 16 February 2011 10:40, Justin wrote:
> I have generated some tasks which call the same URL with different
> parameters. Unfortunately the URL is rate throttled - I can only make
> one call
@David: +1: My exact thoughts! I find that at least once every 2 - 3
weeks GAE will go down like this for hours at a time.
@Everyone: Unless the GAE team improves general uptime, avoid app
engine like the plague! We're only maintaining our last project since
it uses app engine (unfortunate that we
Hi Colin,
We currently do not have plans to allow users to disable the mechanism
we use to lower the latency of tasks.
On 14 February 2011 23:50, Colin Hawkett wrote:
> Thanks Greg - would it be a future possibility to configure the queue
> algorithm not to do this (on a per-queue basis)? Obviou
The problem is now gone. My monitoring logs show that the outage took
place for about 15 minutes from 2011-02-15 18:16:42 to 2011-02-15
18:31:09. In general, I have been having a very good experience with
GAE and this was an unpleasant surprise.
--
You received this message because you are sub
Yep, and now a steady stream of DeadlineExceededExceptions... right on
schedule GAE! Guess I should have expected this; only been a week or
so since this last happened...
I better get to checking my support email, probably some frustrated
customers by now. You think they'd be used to it by this po
+1, just got a sudden burst of over 70 errors in the past 20 minutes.
Normally I just get 1 or 2 every other hour...
On Feb 15, 9:38 pm, PK wrote:
> This started at 2011-02-15 18:16:42.285 PST Anybody else experiencing
> a similar issue right now? Just filed case #4584 Dashboard shows
> that e
This started at 2011-02-15 18:16:42.285 PST Anybody else experiencing
a similar issue right now? Just filed case #4584 Dashboard shows
that everything is fine with the datastore
Thanks
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
There were really way too much error thrown from the datastore for the past
few days. My app is having 8% errors among all requests for the past 18
hours.
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To post to this group, send email
WORKAROUND found!
Set the mimetype for the SVG static files explicitly by including * mime_type:
image/svg+xml *in the .yaml specifications
--
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-appengi
Thanks! I think the problem was that the e-mail I was using is the primary
admin on my Google Apps account. I tried a different e-mail address and it
worked fine. Thanks.
Michael
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To post
Why not just use regular timestamps, and sort descending?
-Nick
On Wed, Feb 16, 2011 at 1:42 AM, Joseph Letness wrote:
> Julian, that is _exactlty_ what I was looking for. Counting up to a
> future time to create descending values... It seems so obvious now!
> It's a much better solution than
just for posterity, can you guys confirm that
register_template_library() worked without any problems in 1.0 & 1.1
(show original snippet?), and only broke with 1.2? (i was under the
impression that it has always taken a module name [rather than a
path].)
thanks!
-- wesley
- - - - - - - - - - - -
I have adapted the source code for mapreduce to add a Cleanup All
function to the UI.
I only modified the code for the python part of mapreduce.
You can find the code and a blog post at
http://djidjadji.appspot.com/2011-02-15-Cleanup-All-for-GAE-mapreduce
--
You received this message because you
Hi Justin,
Have you tried 0.1/s as a rate?
-Nick
On Wed, Feb 16, 2011 at 10:40 AM, Justin wrote:
> I have generated some tasks which call the same URL with different
> parameters. Unfortunately the URL is rate throttled - I can only make
> one call every 10 secs - so I want to configure queue.
I have generated some tasks which call the same URL with different
parameters. Unfortunately the URL is rate throttled - I can only make
one call every 10 secs - so I want to configure queue.yaml to make one
call every 10 secs.
I started by setting the rate parameter at 6/m, but all the tasks were
Hi Jan,
Based on the information you provide its hard to say if that's a
good idea or not; perhaps for your problem it is the best way, maybe
it is not. If you are going to run multiple queries, I would try to
run them asynchronously to reduce the latency.
The idea is not necessarily without
Are yours related to blobstore.FetchData() too? Or a different function?
--
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-appengine@googlegroups.com.
To unsubscribe from this group, send email to
I've been getting random deadline exceed errors too. Don't know what
the deal is... would be nice if there were better ways of debugging
them.
On Feb 15, 2:08 pm, Carlos wrote:
> Suddenly my app started showing this error (after only around 5 seconds of
> processing) when attempting to pull 500k
There is nothing failing ... I guess the good news is the jobs showing as
active are not really running
I tried to delete the entire datastore to duplicate what i can do on the
GoogleAppEngineLauncher via info (clear datastore on launch)
the datastore appears to be deleted but _AE_DatastoreAdmin_
Suddenly my app started showing this error (after only around 5 seconds of
processing) when attempting to pull 500k PDF files from the blobstore, to
serve through a custom handler:
*DeadlineExceededError: The API call blobstore.FetchData() took too long to
respond and was cancelled.
*
Occasion
Directory
com.x.leaders.Directory
Directory
/directory
directory.jsp
The very first time I look at this app each day I receive a 301 Error Too
Many Redirects. After I refresh it will load fine every time afterward. The
jsp is
The gdata java lib is a tricky one. Lots of things can go wrong,
resulting in the same error message. Just a guess: are you adding the
xoauth_requestor_id parameter to the feed URL? It's value should be
the emailaddress of an administrator of the apps domain.
On Feb 14, 4:43 pm, Cherny D. C. Berbe
Hi,
We are thinking of using Google App Engine as our platform to build
custom applications for our clients. As part of our new service
offer, we plan to migrate our clients to Google Apps for Business.
The single cellphone number per Google App Engine account looks like a
problematic limitation
Hi Edward,
(I'm by no means expert, so feel free to discount the following)
In my somewhat limited experience, I've identified only three ways of
making queries faster:
1. denormalizing data (pushing more work into the writes)
2. deferring work (typically via task queues)
3. admitting inco
In our company we feel obstructed by the restriction that a query may
only contain one inequality filter like startsWith.
Currently we are thinking about performing two independent startsWith
queries and combining the result inside the service method.
Do you think this is a good idea?
Does anybody
Ah, Thanks! I'll just switch back to google account authentication. That's
fixed the problem.
Thank you :)
--
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-appengine@googlegroups.com.
To unsubs
hi Nathanael,
Looks like you've got your app set to use federated login (openid).
You need to write a handler for that url (/_ah/login_required) to
setup the login handler.
http://code.google.com/appengine/articles/openid.html
Robert
On Tue, Feb 15, 2011 at 06:51, Nathanael Abbotts wro
ou can also efficiently query
> > over very large periods of time.
> >
> > The key names might look something like:
> >201102151430: 15 Feb 2011 14:30: Journeys that occurred between
> > 14:25 and 14:30.
> >2011021514: 15 Feb 2011 14: Journeys that occ
Julian, that is _exactlty_ what I was looking for. Counting up to a
future time to create descending values... It seems so obvious now!
It's a much better solution than the de-accumulator handler that I
wrote.
Thanks!
On Feb 14, 11:57 pm, Julian Namaro wrote:
> I am not sure about the mathemat
Fantastic! That solved it for me indeed as well. Thanks a million
--
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-appengine@googlegroups.com.
To unsubscribe from this group, send email to
google-
Google fixed this issue last night. Thanks to the Google engineers who
addressed the problem. I'm not sure this would have been addressed if
we hadn't had a friend internal to Google who looked into the issue
for us.
I'm still frustrated that the moderators of this board never replied.
I suppose
Ya it worked for the same way, added a folder with init file and the actual
file inside it.
app_path, taglib = taglib_module.rsplit('.',1)
This is the bug here , trying to split with the last word in the module
path.
for the single word input its raising this kind errors. but this is helpful,
can
Sandeep and Norvo, I was bitten by this bug as well, and I was able to
figure out a solution. It seems that with Django 1.2,
the register_template_library call no longer takes a simple file path.
Instead, it expects a module.class dot-notation. To fix it, I put the file
that contains my custom
@Sandeep
I am encountering the same problem. Did you find a solution?
Thanks
--
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-appengine@googlegroups.com.
To unsubscribe from this group, send emai
To be more clear here are the versions I currently use which I get the
error;
Java(TM) SE Runtime Environment (build 1.6.0_22-b04-307-10M3261)
App Engine API 1.4.2
Google API Client 1.2.2 alpha
With the following versions I'm not getting the same error.
Java(TM) SE Runtime Environment (build 1
Thought some additional details would be helpful:
All other pages on the app respond.
The logs have no errors displayed regarding this, but one warning appears:
/_ah/login_required?continue=http://nat-abbotts.appspot.com/post
1.
No handlers matched this URL.
Changing the versi
Hi Waleed,
Approximately how many requests/minute was your other server making to your
GAE app before it was blocked? Would be interesting to know what the cut-off
point is.
Nick
On 15 February 2011 21:24, Waleed Abdulla wrote:
> Good point. I'll give that a shot.
>
> Regarding the current is
I have an admin page in my app. In the app.yaml, it is configured to be
accessible only to project admins by use of the "login: admin" line.
But when I visit the url that this is set up for, I see this (in Chrome):
Error: Not FoundThe requested URL
/_ah/login_required?continue=http://nat-abbott
I was using 1.2.1 version of google api client library and I updated it to
1.2.2. Its working well with 1.4.0 but now I'm getting a different error for
1.4.2. :)
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To post to this group, send
Hi,
Frew, I'm using JDK 1.6.0 (MacOS X).
Mauro, as far as I know this is not development SDK, I installed it from
Eclipse yesterday. I didn't installed the previous releases of 1.4.2.
Also when I changed the SDK to 1.4.0 from project settings it works.
Thanks.
--
You received this message bec
Good point. I'll give that a shot.
Regarding the current issue, I put my other server on hold for a while and
tried again, and GAE kept rejecting my calls. I tried that a few times,
waiting for more than 10 minutes between tests and it still failed. Finally,
I tried uploading my app to GAE again,
Hi
Maybe you should require auth, we never get the turing bot now that I fixed
the connection
T
--
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-appengine@googlegroups.com.
To unsubscribe from
I'm getting the same issue again today. It's happening more often recently,
which is worrying. I put my other server in maintenance more so it doesn't
call App Engine hoping that after a few minutes GAE would stop blocking my
IP, but so far that's not helping either. I'll try waiting a longer time.
can also efficiently query
> over very large periods of time.
>
> The key names might look something like:
> 201102151430: 15 Feb 2011 14:30: Journeys that occurred between
> 14:25 and 14:30.
> 2011021514: 15 Feb 2011 14: Journeys that occurred from 13:01 to 14:00
>
Copy and pasted from StackOverflow so everyone sees the answer:
It's a really good idea, but: because we're looking for the latest
journey in a unknown time frame, you can't store a how many days old
it is (or equivalent), because if a car stopped making journeys for a
week, the last journey would
We're returning to a web browser. Specifically as JSON.
I'll look into that video!
On Feb 14, 8:35 pm, Calvin wrote:
> Are you returning results to a web browser, or a specialized client? One of
> the Google I/O talks demonstrates spawning a crapload of tasks in parallel
> to collect results, a
53 matches
Mail list logo