You still have to wait.  

>From a code point of view it might look like you are doing an async write, 
not waiting and sending a response back , 
but you will find the response is held until all outstanding rpc's are 
resolved.



On Saturday, December 28, 2013 3:54:22 AM UTC+8, Chad Vincent wrote:
>
> While Vinny probably has the better option, if you wanted something 
> without a rolloff and search options, I would use Objectify and make a 
> LogEntry object.
>
> Just parse the request, build the object, tell Objectify to write it 
> asynchronously, and return.  Should be fast and reliable, but without 
> having to wait for datastore operations to complete before sending the 
> response back.
>
> On Tuesday, December 17, 2013 8:04:24 PM UTC-6, joe.sh...@gmail.com wrote:
>>
>> Hi,
>>
>> I'm after a bit of advice on how best to structure a GAE application.
>>
>> For this application, the frontend instances in GAE will take external 
>> requests, and in turn I need to log strings from all instances to a central 
>> location - ideally just a flat file, as there is no requirement for 
>> NoSQL/SQL and the write operation needs to be as lean/quick as 
>> possible.Storing each line in NoSQL seemed overkill, and the Blob store 
>> also didn't seem appropriate (but perhaps it is?). I'd like to keep the 
>> servlet response time down to an absolute minimum, so had considered the 
>> Task Queue API also (but again, seems like overkill to achieve what I want, 
>> would prefer to handle that off-thread in some way).
>>
>> Initially using 'buckets' in Google Cloud Storage seemed like it might be 
>> appropriate, as the frontend instances can write data to objects there. But 
>> the 'write-once' nature of the bucket files prevents appending to them. I'd 
>> prefer not to have millions of tiny files.
>>
>> At this stage I'm considering a temporary storage location (i.e. in 
>> memory, memcached or NoSQL), and then having a cron job fire every X hours 
>> to grab that data and write a file to GCS, but that feels unwieldy.
>>
>> So I figured I'd ask to make sure I'm not missing something obvious, as 
>> it seems to be a common and basic use case. Ultimately the application may 
>> use map/reduce to take the file/files and post-process them if that helps 
>> in the equation.
>>
>> Cheers,
>> Joe
>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to google-appengine+unsubscr...@googlegroups.com.
To post to this group, send email to google-appengine@googlegroups.com.
Visit this group at http://groups.google.com/group/google-appengine.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to