Hi,

Runnning/Scheduling multiple jobs at once is killing all the other jobs
except only one.

So I have three cubes and to build cube I have corresponding build jobs,
Its failing at third step Build Dimension Dictionary with FileNotFound
exception with

java.io.FileNotFoundException: File does not exist:
/tmp/kylin-65abae6a-72e0-4b59-880b-ece8ab49b33b/sc_sd_esd_diff1/fact_distinct_columns/cn

java.io.FileNotFoundException: File does not exist:
/tmp/kylin-b9145673-de15-4304-8c99-431618219c28/sc_o2s_metrics_verified/fact_distinct_columns/sc

Any suggestions would be highly appreciated!

Thanks,

On Wed, Jun 24, 2015 at 9:32 PM, Vineet Mishra <[email protected]>
wrote:

> Thanks Shi!
>
> On Wed, Jun 24, 2015 at 12:59 PM, Shi, Shaofeng <[email protected]> wrote:
>
>> Yes purge can also be requested via REST API, see the API list:
>>
>>
>> https://github.com/apache/incubator-kylin/blob/master/docs/REST/Kylin%20Res
>> tful%20API%20List.md
>>
>>
>> On 6/24/15, 3:07 PM, "Vineet Mishra" <[email protected]> wrote:
>>
>> >Hi Shi,
>> >
>> >For my use case, its like the data can change throughout from the very
>> >initial for every next day as the hive table is truncate load from
>> scratch
>> >and the process is meant to be like that. I guess in that case purging
>> the
>> >cube and rebuilding from the scratch would be better and only option.
>> >
>> >Referring to the mentioned url for the kylin api
>> >
>> >
>> https://github.com/apache/incubator-kylin/blob/master/docs/REST/Build%20Cu
>> >be%20with%20Restful%20API.md
>> >
>> >is it even possible to purge the cube and rebuild through the api, as I
>> >can
>> >only see build and merge option mentioned in the api.
>> >
>> >Thanks!
>> >
>> >
>> >On Wed, Jun 24, 2015 at 8:13 AM, Shi, Shaofeng <[email protected]> wrote:
>> >
>> >> Hi Vineet,
>> >>
>> >> It can vary depends on your scenario:
>> >>
>> >> Say you have build the data from 23 May (inclusive) to 23 June
>> >> (exclusive); and Now the data of 23 June loads into hive; If the
>> >>historic
>> >> data (23 May to 23 June) in hive will not change, you don’t need to
>> >>build
>> >> that again; You just need build a new date range from 23 to 24 June;
>> >>After
>> >> the build, there will be two cube “segments”: one is for the past
>> month,
>> >> and the second is for the 23 June to 24; we call this as “incremental
>> >> build”; Kylin will scan all cube segments (each segment is a hbase
>> >>table)
>> >> when executing a SQL query, so with a big full build or multiple
>> >> incremental builds you will get the same query result; We suggest use
>> >> incremental build as that will save resource/time on the cube build;
>> >>
>> >> But if your data in hive will change and you expects cube data be sync
>> >> with hive, you need refresh the historic cube segment, or rebuilt the
>> >> whole data range each time;
>> >>
>> >>
>> >> On 6/23/15, 5:40 PM, "Vineet Mishra" <[email protected]> wrote:
>> >>
>> >> >Hi Shi,
>> >> >
>> >> >Referring to the above link, I want to refresh the cube for each day's
>> >> >corresponding last month's data.
>> >> >
>> >> >So my requirement is something like today I wan't to build the cube
>> for
>> >> >last one month data that is from 23 May to 23 June and tomorrow I will
>> >>be
>> >> >requiring the cube for the date range of 24 May to 24 June and so on.
>> >> >
>> >> >Can you shadow me as in that case how to use the mentioned API and
>> move
>> >> >forward? Will it still be considering Refresh build or full cube
>> build.
>> >> >
>> >> >Thanks!
>> >> >
>> >> >On Tue, Jun 23, 2015 at 1:30 AM, Vineet Mishra <
>> [email protected]>
>> >> >wrote:
>> >> >
>> >> >> Hi Shi,
>> >> >>
>> >> >> I am not aware off the basic authentication mechanism mentioned
>> here,
>> >> >> could you help me out as how could I schedule my cube refresh using
>> >>the
>> >> >>API.
>> >> >>
>> >> >> I tried java URL Connection for basic authentication but couldn't
>> get
>> >> >>any
>> >> >> cookies to move further.
>> >> >>
>> >> >> Thanks,
>> >> >>
>> >> >> On Wed, Jun 17, 2015 at 7:21 PM, Shi, Shaofeng <[email protected]>
>> >> wrote:
>> >> >>
>> >> >>> You don¹t need repeatedly create the cube; We call it ³BUILD² or
>> >> >>> ³REFRESH²: ³BUILD² is to build a new segment; ³REFRESH² is to
>> >>update an
>> >> >>> existing cube segment, please check:
>> >> >>>
>> >> >>>
>> >> >>>
>> >> >>>
>> >>
>> https://github.com/apache/incubator-kylin/blob/master/docs/REST/Build%20
>> >> >>>Cub
>> >> >>> e%20with%20Restful%20API.md
>> >> >>>
>> >> >>>
>> >> >>> On 6/11/15, 6:38 PM, "Vineet Mishra" <[email protected]>
>> wrote:
>> >> >>>
>> >> >>> >Hi,
>> >> >>> >
>> >> >>> >Is there a way so that I can schedule the cube creation by purging
>> >>the
>> >> >>> >older cube and creating the new cube everyday taking the same
>> >>window
>> >> >>> >interval say around a month or so.
>> >> >>> >
>> >> >>> >So my requirement is pretty straightforward, I want to build a
>> cube
>> >> >>> >considering for the last one month data set and which should
>> >>refresh
>> >> >>> every
>> >> >>> >day with last one month data from the particular date.
>> >> >>> >
>> >> >>> >Thanks!
>> >> >>>
>> >> >>>
>> >> >>
>> >>
>> >>
>>
>>
>

Reply via email to