predictionio-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From juha.syrj...@gmail.com
Subject Re: Adding a new application is slow
Date Thu, 06 Apr 2017 17:13:30 GMT
I'll try make a patch for this to PredicitionIO during the weekend if I 
have the time.

On Thursday, April 6, 2017 at 8:07:31 PM UTC+3, pat wrote:
>
> Hmm, interesting. I have heard of this being a problem in PredictionIO but 
> haven’t heard this explanation before IIRC.
>
>
> On Apr 6, 2017, at 9:47 AM, juha.s...@gmail.com <javascript:> wrote:
>
> Ah, I just found the problem: a fresh AWS instance takes a long time to 
> generate enough entropy so that a random access key can be generated. 
>
> Related github issue: 
> https://github.com/apache/incubator-predictionio/issues/253
>
> More discussion about the a similar problem: 
> https://groups.google.com/forum/#!searchin/actionml-user/aws%7Csort:relevance/actionml-user/N9o2jBkyF_w/bUYr5U8zFQAJ
>
>
> On Thursday, April 6, 2017 at 7:33:49 PM UTC+3, juha.s...@gmail.com wrote:
>>
>> Hi,
>>
>> Yes, it is clear that the AWS image is for dev only. I was just wondering 
>> the slowness of the a newly created instance when it has no data. 
>>
>> However can't be that the HDFS is out of space. I ran those commands 
>> right after creating a new AWS instance. I don't have any engines running 
>> or configured, and I have not added any data yet to the system. The machine 
>> seems to be completely idle while adding the new application.
>>
>> Current HDFS usage:
>>
>> $ ./hadoop fs -df -h
>>
>> Filesystem              Size   Used  Available  Use%
>>
>> hdfs://127.0.0.1:9000  7.7 G  220 K      1.9 G    0%
>>
>> Also system has plenty of free memory
>>
>> $ free -h
>>
>>               total        used        free      shared  buff/cache   
>> available
>>
>> Mem:            14G        2.0G         12G        8.6M        404M       
>>   12G
>>
>> Swap:            0B          0B          0B
>>
>>
>>
>> On Thursday, April 6, 2017 at 6:47:53 PM UTC+3, pat wrote:
>>>
>>> Yes, this is sometimes true and it could be a couple things. Some 
>>> operations in HDFS are slow to complete since they require replication. 
>>> Even starting a cluster can take minutes. Is it a problem? Is it possible 
>>> HBase is getting full (check remaining data in HDFS). or needs to be 
>>> scaled? The all-in-one instance on AWS only uses vertical scaling by number 
>>> of cores, amount of memory, and disk space since all services run on one 
>>> machine. You can “stop” the instance and change the type to a larger 
>>> instance then “start” it to get it scaled. You have to snapshot the disk
to 
>>> scale disk by copying to a new one and attaching the new volume.
>>>
>>> I hope it was clear that the AWS AMI was meant for dev, not production.
>>>
>>> On Apr 6, 2017, at 6:12 AM, juha.s...@gmail.com wrote:
>>>
>>> Hi,
>>>
>>> I am using ready-made AWS image from 
>>> http://actionml.com/docs/awssetupguide with r3.large instance in 
>>> eu-central-1 region.
>>>
>>> I followed installation instructions to the letter. Integration test 
>>> passes but takes many minutes. Probably because creating a new a new app 
>>> takes so long.
>>>
>>> Then I tried creating a new application and that takes about 10 minutes. 
>>> Is this normal? 
>>>
>>> What this command is actually doing? I had assumed that it would just 
>>> create few records to the databases, which should take just a dozen seconds 
>>> or so.
>>>
>>> Here is a log for creating a new application. Most of the time is spent 
>>> lines 'HBLEvents' and '[App$] Initialized Event...'
>>>
>>> $ time pio app new foobar123
>>> [INFO] [HBLEvents] The table pio_event:events_3 doesn't exist yet. 
>>> Creating now...
>>> [INFO] [App$] Initialized Event Store for this app ID: 3.
>>> [INFO] [App$] Created new app:
>>> [INFO] [App$]       Name: foobar123
>>> [INFO] [App$]         ID: 3
>>> [INFO] [App$] Access Key: ....
>>>
>>> real 9m58.891s
>>> user 0m9.632s
>>> sys 0m0.352s
>>>
>>>
>>> -- 
>>> Juha Syrjälä
>>>
>>> -- 
>>> You received this message because you are subscribed to the Google 
>>> Groups "actionml-user" group.
>>> To unsubscribe from this group and stop receiving emails from it, send 
>>> an email to actionml-use...@googlegroups.com.
>>> To post to this group, send email to action...@googlegroups.com.
>>> To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/actionml-user/6a38952f-6f23-445f-bea8-e97fbc04aa4b%40googlegroups.com

>>> <https://groups.google.com/d/msgid/actionml-user/6a38952f-6f23-445f-bea8-e97fbc04aa4b%40googlegroups.com?utm_medium=email&utm_source=footer>
>>> .
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>>
> -- 
> You received this message because you are subscribed to the Google Groups 
> "actionml-user" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to actionml-use...@googlegroups.com <javascript:>.
> To post to this group, send email to action...@googlegroups.com 
> <javascript:>.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/actionml-user/76ce9fa7-c6b0-4695-8b2f-724dee38653a%40googlegroups.com

> <https://groups.google.com/d/msgid/actionml-user/76ce9fa7-c6b0-4695-8b2f-724dee38653a%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
> For more options, visit https://groups.google.com/d/optout.
>
>
Mime
  • Unnamed multipart/mixed (inline, None, 0 bytes)
View raw message