You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kylin.apache.org by Jie Tao <ji...@gameforge.com> on 2016/06/14 10:57:53 UTC
kylin intermediate tables in Hive
Kylin actually drops useless intermediate tables after cube building,
but I still see one "kylin_intermediate_cubename_searchdata" table for
each cube building in Hive. Are these tables still usefull for Kylin? I
use Kylin 1.5.2.1.
Cheers,
Jie
Re: kylin intermediate tables in Hive
Posted by hongbin ma <ma...@apache.org>.
it might have something to do with the bug described in
https://issues.apache.org/jira/browse/KYLIN-1828, I'll cross check
On Tue, Jun 21, 2016 at 10:36 PM, ShaoFeng Shi <sh...@apache.org>
wrote:
> Hi Jie, would you mind to report a JIRA with this problem you found? If you
> can attache the json of this job and a couple of screen shot, that would be
> great for analysis. Thank you!
>
> 2016-06-21 16:49 GMT+08:00 Jie Tao <ji...@gameforge.com>:
>
> > actually the jobs have been disgarded. Maybe it is a bug that the status
> > of the job is still "ERROR". I have a look of the jobs at the Kylin Web
> UI
> > and found that the job was not marked with black colors like other
> > discarded jobs althouth the "action" button only has one choice
> > "diagonose". My Kylin is 1.5.2.1.
> >
> > Cheers,
> >
> > Jie
> >
> >
> > Am 17.06.2016 um 11:05 schrieb ShaoFeng Shi:
> >
> >> by default the web UI only shows the jobs in LAST ONE WEEK, pls have a
> >> check.
> >>
> >> 2016-06-17 16:58 GMT+08:00 Jie Tao <ji...@gameforge.com>:
> >>
> >> actually I discarded all jobs and I do not see any ERROR job in the
> >>> Monitor view of Kylin UI.
> >>>
> >>> Where can I see these error jobs?
> >>>
> >>> Jie
> >>>
> >>>
> >>> Am 17.06.2016 um 10:31 schrieb ShaoFeng Shi:
> >>>
> >>> Hi Jie,
> >>>>
> >>>> If a job is "ERROR", the intermediate hive table of it will not be
> >>>> dropped,
> >>>> as "ERROR" is not a final state; User can resume an "Error" job at any
> >>>> time, so Kylin skipped to cleanup for that.
> >>>>
> >>>> If you discard these error jobs, and re-run the cleanup, the
> >>>> intermediate
> >>>> hive table will be dropped.
> >>>>
> >>>> The message here is not clear, will change the wording...
> >>>>
> >>>> 2016-06-17 15:48 GMT+08:00 Jie Tao <ji...@gameforge.com>:
> >>>>
> >>>> You are correct, the intermediate tables are left by fail-building. I
> do
> >>>>
> >>>>> clean up storage based on the linked guide. Intermediate data in HDFS
> >>>>> and
> >>>>> Hbase are deleted, but the intermediate tables in Hive not. The
> command
> >>>>> shows the tables but do not drop them. I donot have a lookup table
> but
> >>>>> my
> >>>>> fact table is a view.
> >>>>>
> >>>>> As I run the cleanup command,
> >>>>> kylin_intermediate_logout_full_cube_19700101000000_20151001000000
> >>>>> kylin_intermediate_logout_full_cube_19700101000000_20160529010500
> >>>>> kylin_intermediate_logout_full_cube_19700101000000_20160608000000
> >>>>> kylin_intermediate_logout_full_cube_19700101000000_20160608010500
> >>>>> kylin_intermediate_logout_full_cube_19700101000000_20160609010500
> >>>>> kylin_intermediate_logout_full_cube_19700101000000_20160615000000
> >>>>> kylin_intermediate_logout_full_cube_19700101000000_20160626000000
> >>>>> kylin_intermediate_logout_full_cube_19700101000000_20160626042000
> >>>>> kylin_intermediate_test_cube_19700101000000_20151201010500
> >>>>> kylin_intermediate_test_cube_19700101000000_20151231234000
> >>>>> kylin_intermediate_test_cube_19700101000000_20160302063000
> >>>>> kylin_intermediate_test_cube_19700101000000_20160626000000
> >>>>> kylin_intermediate_test_cube_19700101000000_20160626042000
> >>>>> kylin_intermediate_test_cube_19700101000000_20160704082000
> >>>>> Time taken: 0.189 seconds, Fetched: 14 row(s)
> >>>>> 2016-06-17 09:37:12,645 INFO [main StorageCleanupJob:262]: Remove
> >>>>> intermediate hive table with job id
> >>>>> 493fd20b-3074-403e-9963-fe4fb7ff7c65
> >>>>> with job status ERROR
> >>>>> 2016-06-17 09:37:12,648 INFO [main StorageCleanupJob:262]: Remove
> >>>>> intermediate hive table with job id
> >>>>> 8a377e30-e3ba-4fe2-be12-e7d412afec5e
> >>>>> with job status ERROR
> >>>>>
> >>>>> Best regards,
> >>>>>
> >>>>> Jie
> >>>>>
> >>>>>
> >>>>> Am 17.06.2016 um 09:16 schrieb ShaoFeng Shi:
> >>>>>
> >>>>> BTW, are you using a view as lookup table?
> >>>>>
> >>>>>> 2016-06-17 15:15 GMT+08:00 ShaoFeng Shi <sh...@apache.org>:
> >>>>>>
> >>>>>> This is common; If you have a job failed in between, and you discard
> >>>>>> that
> >>>>>>
> >>>>>> job, the "Garbage collection" step will not be executed, so the
> >>>>>>> garbages
> >>>>>>> will be left there.
> >>>>>>>
> >>>>>>> This is why we still recommend user to run offline cleanup every
> some
> >>>>>>> period; It is not perfert, but be good for most scenarios:
> >>>>>>> https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
> >>>>>>>
> >>>>>>> 2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
> >>>>>>>
> >>>>>>> Woo... something new to me. Anybody knows?
> >>>>>>>
> >>>>>>> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com>
> >>>>>>>> wrote:
> >>>>>>>>
> >>>>>>>> Kylin actually drops useless intermediate tables after cube
> >>>>>>>> building,
> >>>>>>>> but
> >>>>>>>>
> >>>>>>>> I still see one "kylin_intermediate_cubename_searchdata" table for
> >>>>>>>>
> >>>>>>>>> each
> >>>>>>>>> cube building in Hive. Are these tables still usefull for Kylin?
> I
> >>>>>>>>> use
> >>>>>>>>> Kylin 1.5.2.1.
> >>>>>>>>>
> >>>>>>>>> Cheers,
> >>>>>>>>>
> >>>>>>>>> Jie
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> --
> >>>>>>>>>
> >>>>>>>> Best regards,
> >>>>>>>
> >>>>>>> Shaofeng Shi
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>
> >
>
>
> --
> Best regards,
>
> Shaofeng Shi
>
--
Regards,
*Bin Mahone | 马洪宾*
Re: kylin intermediate tables in Hive
Posted by ShaoFeng Shi <sh...@apache.org>.
Hi Jie, would you mind to report a JIRA with this problem you found? If you
can attache the json of this job and a couple of screen shot, that would be
great for analysis. Thank you!
2016-06-21 16:49 GMT+08:00 Jie Tao <ji...@gameforge.com>:
> actually the jobs have been disgarded. Maybe it is a bug that the status
> of the job is still "ERROR". I have a look of the jobs at the Kylin Web UI
> and found that the job was not marked with black colors like other
> discarded jobs althouth the "action" button only has one choice
> "diagonose". My Kylin is 1.5.2.1.
>
> Cheers,
>
> Jie
>
>
> Am 17.06.2016 um 11:05 schrieb ShaoFeng Shi:
>
>> by default the web UI only shows the jobs in LAST ONE WEEK, pls have a
>> check.
>>
>> 2016-06-17 16:58 GMT+08:00 Jie Tao <ji...@gameforge.com>:
>>
>> actually I discarded all jobs and I do not see any ERROR job in the
>>> Monitor view of Kylin UI.
>>>
>>> Where can I see these error jobs?
>>>
>>> Jie
>>>
>>>
>>> Am 17.06.2016 um 10:31 schrieb ShaoFeng Shi:
>>>
>>> Hi Jie,
>>>>
>>>> If a job is "ERROR", the intermediate hive table of it will not be
>>>> dropped,
>>>> as "ERROR" is not a final state; User can resume an "Error" job at any
>>>> time, so Kylin skipped to cleanup for that.
>>>>
>>>> If you discard these error jobs, and re-run the cleanup, the
>>>> intermediate
>>>> hive table will be dropped.
>>>>
>>>> The message here is not clear, will change the wording...
>>>>
>>>> 2016-06-17 15:48 GMT+08:00 Jie Tao <ji...@gameforge.com>:
>>>>
>>>> You are correct, the intermediate tables are left by fail-building. I do
>>>>
>>>>> clean up storage based on the linked guide. Intermediate data in HDFS
>>>>> and
>>>>> Hbase are deleted, but the intermediate tables in Hive not. The command
>>>>> shows the tables but do not drop them. I donot have a lookup table but
>>>>> my
>>>>> fact table is a view.
>>>>>
>>>>> As I run the cleanup command,
>>>>> kylin_intermediate_logout_full_cube_19700101000000_20151001000000
>>>>> kylin_intermediate_logout_full_cube_19700101000000_20160529010500
>>>>> kylin_intermediate_logout_full_cube_19700101000000_20160608000000
>>>>> kylin_intermediate_logout_full_cube_19700101000000_20160608010500
>>>>> kylin_intermediate_logout_full_cube_19700101000000_20160609010500
>>>>> kylin_intermediate_logout_full_cube_19700101000000_20160615000000
>>>>> kylin_intermediate_logout_full_cube_19700101000000_20160626000000
>>>>> kylin_intermediate_logout_full_cube_19700101000000_20160626042000
>>>>> kylin_intermediate_test_cube_19700101000000_20151201010500
>>>>> kylin_intermediate_test_cube_19700101000000_20151231234000
>>>>> kylin_intermediate_test_cube_19700101000000_20160302063000
>>>>> kylin_intermediate_test_cube_19700101000000_20160626000000
>>>>> kylin_intermediate_test_cube_19700101000000_20160626042000
>>>>> kylin_intermediate_test_cube_19700101000000_20160704082000
>>>>> Time taken: 0.189 seconds, Fetched: 14 row(s)
>>>>> 2016-06-17 09:37:12,645 INFO [main StorageCleanupJob:262]: Remove
>>>>> intermediate hive table with job id
>>>>> 493fd20b-3074-403e-9963-fe4fb7ff7c65
>>>>> with job status ERROR
>>>>> 2016-06-17 09:37:12,648 INFO [main StorageCleanupJob:262]: Remove
>>>>> intermediate hive table with job id
>>>>> 8a377e30-e3ba-4fe2-be12-e7d412afec5e
>>>>> with job status ERROR
>>>>>
>>>>> Best regards,
>>>>>
>>>>> Jie
>>>>>
>>>>>
>>>>> Am 17.06.2016 um 09:16 schrieb ShaoFeng Shi:
>>>>>
>>>>> BTW, are you using a view as lookup table?
>>>>>
>>>>>> 2016-06-17 15:15 GMT+08:00 ShaoFeng Shi <sh...@apache.org>:
>>>>>>
>>>>>> This is common; If you have a job failed in between, and you discard
>>>>>> that
>>>>>>
>>>>>> job, the "Garbage collection" step will not be executed, so the
>>>>>>> garbages
>>>>>>> will be left there.
>>>>>>>
>>>>>>> This is why we still recommend user to run offline cleanup every some
>>>>>>> period; It is not perfert, but be good for most scenarios:
>>>>>>> https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
>>>>>>>
>>>>>>> 2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
>>>>>>>
>>>>>>> Woo... something new to me. Anybody knows?
>>>>>>>
>>>>>>> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>> Kylin actually drops useless intermediate tables after cube
>>>>>>>> building,
>>>>>>>> but
>>>>>>>>
>>>>>>>> I still see one "kylin_intermediate_cubename_searchdata" table for
>>>>>>>>
>>>>>>>>> each
>>>>>>>>> cube building in Hive. Are these tables still usefull for Kylin? I
>>>>>>>>> use
>>>>>>>>> Kylin 1.5.2.1.
>>>>>>>>>
>>>>>>>>> Cheers,
>>>>>>>>>
>>>>>>>>> Jie
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>>
>>>>>>>> Best regards,
>>>>>>>
>>>>>>> Shaofeng Shi
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>
>
--
Best regards,
Shaofeng Shi
Re: kylin intermediate tables in Hive
Posted by Jie Tao <ji...@gameforge.com>.
actually the jobs have been disgarded. Maybe it is a bug that the status
of the job is still "ERROR". I have a look of the jobs at the Kylin Web
UI and found that the job was not marked with black colors like other
discarded jobs althouth the "action" button only has one choice
"diagonose". My Kylin is 1.5.2.1.
Cheers,
Jie
Am 17.06.2016 um 11:05 schrieb ShaoFeng Shi:
> by default the web UI only shows the jobs in LAST ONE WEEK, pls have a
> check.
>
> 2016-06-17 16:58 GMT+08:00 Jie Tao <ji...@gameforge.com>:
>
>> actually I discarded all jobs and I do not see any ERROR job in the
>> Monitor view of Kylin UI.
>>
>> Where can I see these error jobs?
>>
>> Jie
>>
>>
>> Am 17.06.2016 um 10:31 schrieb ShaoFeng Shi:
>>
>>> Hi Jie,
>>>
>>> If a job is "ERROR", the intermediate hive table of it will not be
>>> dropped,
>>> as "ERROR" is not a final state; User can resume an "Error" job at any
>>> time, so Kylin skipped to cleanup for that.
>>>
>>> If you discard these error jobs, and re-run the cleanup, the intermediate
>>> hive table will be dropped.
>>>
>>> The message here is not clear, will change the wording...
>>>
>>> 2016-06-17 15:48 GMT+08:00 Jie Tao <ji...@gameforge.com>:
>>>
>>> You are correct, the intermediate tables are left by fail-building. I do
>>>> clean up storage based on the linked guide. Intermediate data in HDFS and
>>>> Hbase are deleted, but the intermediate tables in Hive not. The command
>>>> shows the tables but do not drop them. I donot have a lookup table but my
>>>> fact table is a view.
>>>>
>>>> As I run the cleanup command,
>>>> kylin_intermediate_logout_full_cube_19700101000000_20151001000000
>>>> kylin_intermediate_logout_full_cube_19700101000000_20160529010500
>>>> kylin_intermediate_logout_full_cube_19700101000000_20160608000000
>>>> kylin_intermediate_logout_full_cube_19700101000000_20160608010500
>>>> kylin_intermediate_logout_full_cube_19700101000000_20160609010500
>>>> kylin_intermediate_logout_full_cube_19700101000000_20160615000000
>>>> kylin_intermediate_logout_full_cube_19700101000000_20160626000000
>>>> kylin_intermediate_logout_full_cube_19700101000000_20160626042000
>>>> kylin_intermediate_test_cube_19700101000000_20151201010500
>>>> kylin_intermediate_test_cube_19700101000000_20151231234000
>>>> kylin_intermediate_test_cube_19700101000000_20160302063000
>>>> kylin_intermediate_test_cube_19700101000000_20160626000000
>>>> kylin_intermediate_test_cube_19700101000000_20160626042000
>>>> kylin_intermediate_test_cube_19700101000000_20160704082000
>>>> Time taken: 0.189 seconds, Fetched: 14 row(s)
>>>> 2016-06-17 09:37:12,645 INFO [main StorageCleanupJob:262]: Remove
>>>> intermediate hive table with job id 493fd20b-3074-403e-9963-fe4fb7ff7c65
>>>> with job status ERROR
>>>> 2016-06-17 09:37:12,648 INFO [main StorageCleanupJob:262]: Remove
>>>> intermediate hive table with job id 8a377e30-e3ba-4fe2-be12-e7d412afec5e
>>>> with job status ERROR
>>>>
>>>> Best regards,
>>>>
>>>> Jie
>>>>
>>>>
>>>> Am 17.06.2016 um 09:16 schrieb ShaoFeng Shi:
>>>>
>>>> BTW, are you using a view as lookup table?
>>>>> 2016-06-17 15:15 GMT+08:00 ShaoFeng Shi <sh...@apache.org>:
>>>>>
>>>>> This is common; If you have a job failed in between, and you discard
>>>>> that
>>>>>
>>>>>> job, the "Garbage collection" step will not be executed, so the
>>>>>> garbages
>>>>>> will be left there.
>>>>>>
>>>>>> This is why we still recommend user to run offline cleanup every some
>>>>>> period; It is not perfert, but be good for most scenarios:
>>>>>> https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
>>>>>>
>>>>>> 2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
>>>>>>
>>>>>> Woo... something new to me. Anybody knows?
>>>>>>
>>>>>>> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com>
>>>>>>> wrote:
>>>>>>>
>>>>>>> Kylin actually drops useless intermediate tables after cube building,
>>>>>>> but
>>>>>>>
>>>>>>> I still see one "kylin_intermediate_cubename_searchdata" table for
>>>>>>>> each
>>>>>>>> cube building in Hive. Are these tables still usefull for Kylin? I
>>>>>>>> use
>>>>>>>> Kylin 1.5.2.1.
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>>
>>>>>>>> Jie
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>> Best regards,
>>>>>>
>>>>>> Shaofeng Shi
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>
Re: kylin intermediate tables in Hive
Posted by ShaoFeng Shi <sh...@apache.org>.
by default the web UI only shows the jobs in LAST ONE WEEK, pls have a
check.
2016-06-17 16:58 GMT+08:00 Jie Tao <ji...@gameforge.com>:
> actually I discarded all jobs and I do not see any ERROR job in the
> Monitor view of Kylin UI.
>
> Where can I see these error jobs?
>
> Jie
>
>
> Am 17.06.2016 um 10:31 schrieb ShaoFeng Shi:
>
>> Hi Jie,
>>
>> If a job is "ERROR", the intermediate hive table of it will not be
>> dropped,
>> as "ERROR" is not a final state; User can resume an "Error" job at any
>> time, so Kylin skipped to cleanup for that.
>>
>> If you discard these error jobs, and re-run the cleanup, the intermediate
>> hive table will be dropped.
>>
>> The message here is not clear, will change the wording...
>>
>> 2016-06-17 15:48 GMT+08:00 Jie Tao <ji...@gameforge.com>:
>>
>> You are correct, the intermediate tables are left by fail-building. I do
>>> clean up storage based on the linked guide. Intermediate data in HDFS and
>>> Hbase are deleted, but the intermediate tables in Hive not. The command
>>> shows the tables but do not drop them. I donot have a lookup table but my
>>> fact table is a view.
>>>
>>> As I run the cleanup command,
>>> kylin_intermediate_logout_full_cube_19700101000000_20151001000000
>>> kylin_intermediate_logout_full_cube_19700101000000_20160529010500
>>> kylin_intermediate_logout_full_cube_19700101000000_20160608000000
>>> kylin_intermediate_logout_full_cube_19700101000000_20160608010500
>>> kylin_intermediate_logout_full_cube_19700101000000_20160609010500
>>> kylin_intermediate_logout_full_cube_19700101000000_20160615000000
>>> kylin_intermediate_logout_full_cube_19700101000000_20160626000000
>>> kylin_intermediate_logout_full_cube_19700101000000_20160626042000
>>> kylin_intermediate_test_cube_19700101000000_20151201010500
>>> kylin_intermediate_test_cube_19700101000000_20151231234000
>>> kylin_intermediate_test_cube_19700101000000_20160302063000
>>> kylin_intermediate_test_cube_19700101000000_20160626000000
>>> kylin_intermediate_test_cube_19700101000000_20160626042000
>>> kylin_intermediate_test_cube_19700101000000_20160704082000
>>> Time taken: 0.189 seconds, Fetched: 14 row(s)
>>> 2016-06-17 09:37:12,645 INFO [main StorageCleanupJob:262]: Remove
>>> intermediate hive table with job id 493fd20b-3074-403e-9963-fe4fb7ff7c65
>>> with job status ERROR
>>> 2016-06-17 09:37:12,648 INFO [main StorageCleanupJob:262]: Remove
>>> intermediate hive table with job id 8a377e30-e3ba-4fe2-be12-e7d412afec5e
>>> with job status ERROR
>>>
>>> Best regards,
>>>
>>> Jie
>>>
>>>
>>> Am 17.06.2016 um 09:16 schrieb ShaoFeng Shi:
>>>
>>> BTW, are you using a view as lookup table?
>>>>
>>>> 2016-06-17 15:15 GMT+08:00 ShaoFeng Shi <sh...@apache.org>:
>>>>
>>>> This is common; If you have a job failed in between, and you discard
>>>> that
>>>>
>>>>> job, the "Garbage collection" step will not be executed, so the
>>>>> garbages
>>>>> will be left there.
>>>>>
>>>>> This is why we still recommend user to run offline cleanup every some
>>>>> period; It is not perfert, but be good for most scenarios:
>>>>> https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
>>>>>
>>>>> 2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
>>>>>
>>>>> Woo... something new to me. Anybody knows?
>>>>>
>>>>>> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com>
>>>>>> wrote:
>>>>>>
>>>>>> Kylin actually drops useless intermediate tables after cube building,
>>>>>> but
>>>>>>
>>>>>> I still see one "kylin_intermediate_cubename_searchdata" table for
>>>>>>> each
>>>>>>> cube building in Hive. Are these tables still usefull for Kylin? I
>>>>>>> use
>>>>>>> Kylin 1.5.2.1.
>>>>>>>
>>>>>>> Cheers,
>>>>>>>
>>>>>>> Jie
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>> Best regards,
>>>>>
>>>>> Shaofeng Shi
>>>>>
>>>>>
>>>>>
>>>>>
>>
>
--
Best regards,
Shaofeng Shi
Re: kylin intermediate tables in Hive
Posted by Jie Tao <ji...@gameforge.com>.
actually I discarded all jobs and I do not see any ERROR job in the
Monitor view of Kylin UI.
Where can I see these error jobs?
Jie
Am 17.06.2016 um 10:31 schrieb ShaoFeng Shi:
> Hi Jie,
>
> If a job is "ERROR", the intermediate hive table of it will not be dropped,
> as "ERROR" is not a final state; User can resume an "Error" job at any
> time, so Kylin skipped to cleanup for that.
>
> If you discard these error jobs, and re-run the cleanup, the intermediate
> hive table will be dropped.
>
> The message here is not clear, will change the wording...
>
> 2016-06-17 15:48 GMT+08:00 Jie Tao <ji...@gameforge.com>:
>
>> You are correct, the intermediate tables are left by fail-building. I do
>> clean up storage based on the linked guide. Intermediate data in HDFS and
>> Hbase are deleted, but the intermediate tables in Hive not. The command
>> shows the tables but do not drop them. I donot have a lookup table but my
>> fact table is a view.
>>
>> As I run the cleanup command,
>> kylin_intermediate_logout_full_cube_19700101000000_20151001000000
>> kylin_intermediate_logout_full_cube_19700101000000_20160529010500
>> kylin_intermediate_logout_full_cube_19700101000000_20160608000000
>> kylin_intermediate_logout_full_cube_19700101000000_20160608010500
>> kylin_intermediate_logout_full_cube_19700101000000_20160609010500
>> kylin_intermediate_logout_full_cube_19700101000000_20160615000000
>> kylin_intermediate_logout_full_cube_19700101000000_20160626000000
>> kylin_intermediate_logout_full_cube_19700101000000_20160626042000
>> kylin_intermediate_test_cube_19700101000000_20151201010500
>> kylin_intermediate_test_cube_19700101000000_20151231234000
>> kylin_intermediate_test_cube_19700101000000_20160302063000
>> kylin_intermediate_test_cube_19700101000000_20160626000000
>> kylin_intermediate_test_cube_19700101000000_20160626042000
>> kylin_intermediate_test_cube_19700101000000_20160704082000
>> Time taken: 0.189 seconds, Fetched: 14 row(s)
>> 2016-06-17 09:37:12,645 INFO [main StorageCleanupJob:262]: Remove
>> intermediate hive table with job id 493fd20b-3074-403e-9963-fe4fb7ff7c65
>> with job status ERROR
>> 2016-06-17 09:37:12,648 INFO [main StorageCleanupJob:262]: Remove
>> intermediate hive table with job id 8a377e30-e3ba-4fe2-be12-e7d412afec5e
>> with job status ERROR
>>
>> Best regards,
>>
>> Jie
>>
>>
>> Am 17.06.2016 um 09:16 schrieb ShaoFeng Shi:
>>
>>> BTW, are you using a view as lookup table?
>>>
>>> 2016-06-17 15:15 GMT+08:00 ShaoFeng Shi <sh...@apache.org>:
>>>
>>> This is common; If you have a job failed in between, and you discard that
>>>> job, the "Garbage collection" step will not be executed, so the garbages
>>>> will be left there.
>>>>
>>>> This is why we still recommend user to run offline cleanup every some
>>>> period; It is not perfert, but be good for most scenarios:
>>>> https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
>>>>
>>>> 2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
>>>>
>>>> Woo... something new to me. Anybody knows?
>>>>> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com> wrote:
>>>>>
>>>>> Kylin actually drops useless intermediate tables after cube building,
>>>>> but
>>>>>
>>>>>> I still see one "kylin_intermediate_cubename_searchdata" table for each
>>>>>> cube building in Hive. Are these tables still usefull for Kylin? I use
>>>>>> Kylin 1.5.2.1.
>>>>>>
>>>>>> Cheers,
>>>>>>
>>>>>> Jie
>>>>>>
>>>>>>
>>>> --
>>>> Best regards,
>>>>
>>>> Shaofeng Shi
>>>>
>>>>
>>>>
>
Re: kylin intermediate tables in Hive
Posted by ShaoFeng Shi <sh...@apache.org>.
Hi Jie,
If a job is "ERROR", the intermediate hive table of it will not be dropped,
as "ERROR" is not a final state; User can resume an "Error" job at any
time, so Kylin skipped to cleanup for that.
If you discard these error jobs, and re-run the cleanup, the intermediate
hive table will be dropped.
The message here is not clear, will change the wording...
2016-06-17 15:48 GMT+08:00 Jie Tao <ji...@gameforge.com>:
> You are correct, the intermediate tables are left by fail-building. I do
> clean up storage based on the linked guide. Intermediate data in HDFS and
> Hbase are deleted, but the intermediate tables in Hive not. The command
> shows the tables but do not drop them. I donot have a lookup table but my
> fact table is a view.
>
> As I run the cleanup command,
> kylin_intermediate_logout_full_cube_19700101000000_20151001000000
> kylin_intermediate_logout_full_cube_19700101000000_20160529010500
> kylin_intermediate_logout_full_cube_19700101000000_20160608000000
> kylin_intermediate_logout_full_cube_19700101000000_20160608010500
> kylin_intermediate_logout_full_cube_19700101000000_20160609010500
> kylin_intermediate_logout_full_cube_19700101000000_20160615000000
> kylin_intermediate_logout_full_cube_19700101000000_20160626000000
> kylin_intermediate_logout_full_cube_19700101000000_20160626042000
> kylin_intermediate_test_cube_19700101000000_20151201010500
> kylin_intermediate_test_cube_19700101000000_20151231234000
> kylin_intermediate_test_cube_19700101000000_20160302063000
> kylin_intermediate_test_cube_19700101000000_20160626000000
> kylin_intermediate_test_cube_19700101000000_20160626042000
> kylin_intermediate_test_cube_19700101000000_20160704082000
> Time taken: 0.189 seconds, Fetched: 14 row(s)
> 2016-06-17 09:37:12,645 INFO [main StorageCleanupJob:262]: Remove
> intermediate hive table with job id 493fd20b-3074-403e-9963-fe4fb7ff7c65
> with job status ERROR
> 2016-06-17 09:37:12,648 INFO [main StorageCleanupJob:262]: Remove
> intermediate hive table with job id 8a377e30-e3ba-4fe2-be12-e7d412afec5e
> with job status ERROR
>
> Best regards,
>
> Jie
>
>
> Am 17.06.2016 um 09:16 schrieb ShaoFeng Shi:
>
>> BTW, are you using a view as lookup table?
>>
>> 2016-06-17 15:15 GMT+08:00 ShaoFeng Shi <sh...@apache.org>:
>>
>> This is common; If you have a job failed in between, and you discard that
>>> job, the "Garbage collection" step will not be executed, so the garbages
>>> will be left there.
>>>
>>> This is why we still recommend user to run offline cleanup every some
>>> period; It is not perfert, but be good for most scenarios:
>>> https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
>>>
>>> 2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
>>>
>>> Woo... something new to me. Anybody knows?
>>>>
>>>> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com> wrote:
>>>>
>>>> Kylin actually drops useless intermediate tables after cube building,
>>>>>
>>>> but
>>>>
>>>>> I still see one "kylin_intermediate_cubename_searchdata" table for each
>>>>> cube building in Hive. Are these tables still usefull for Kylin? I use
>>>>> Kylin 1.5.2.1.
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Jie
>>>>>
>>>>>
>>>
>>> --
>>> Best regards,
>>>
>>> Shaofeng Shi
>>>
>>>
>>>
>>
>
--
Best regards,
Shaofeng Shi
Re: kylin intermediate tables in Hive
Posted by Jie Tao <ji...@gameforge.com>.
You are correct, the intermediate tables are left by fail-building. I do
clean up storage based on the linked guide. Intermediate data in HDFS
and Hbase are deleted, but the intermediate tables in Hive not. The
command shows the tables but do not drop them. I donot have a lookup
table but my fact table is a view.
As I run the cleanup command,
kylin_intermediate_logout_full_cube_19700101000000_20151001000000
kylin_intermediate_logout_full_cube_19700101000000_20160529010500
kylin_intermediate_logout_full_cube_19700101000000_20160608000000
kylin_intermediate_logout_full_cube_19700101000000_20160608010500
kylin_intermediate_logout_full_cube_19700101000000_20160609010500
kylin_intermediate_logout_full_cube_19700101000000_20160615000000
kylin_intermediate_logout_full_cube_19700101000000_20160626000000
kylin_intermediate_logout_full_cube_19700101000000_20160626042000
kylin_intermediate_test_cube_19700101000000_20151201010500
kylin_intermediate_test_cube_19700101000000_20151231234000
kylin_intermediate_test_cube_19700101000000_20160302063000
kylin_intermediate_test_cube_19700101000000_20160626000000
kylin_intermediate_test_cube_19700101000000_20160626042000
kylin_intermediate_test_cube_19700101000000_20160704082000
Time taken: 0.189 seconds, Fetched: 14 row(s)
2016-06-17 09:37:12,645 INFO [main StorageCleanupJob:262]: Remove
intermediate hive table with job id 493fd20b-3074-403e-9963-fe4fb7ff7c65
with job status ERROR
2016-06-17 09:37:12,648 INFO [main StorageCleanupJob:262]: Remove
intermediate hive table with job id 8a377e30-e3ba-4fe2-be12-e7d412afec5e
with job status ERROR
Best regards,
Jie
Am 17.06.2016 um 09:16 schrieb ShaoFeng Shi:
> BTW, are you using a view as lookup table?
>
> 2016-06-17 15:15 GMT+08:00 ShaoFeng Shi <sh...@apache.org>:
>
>> This is common; If you have a job failed in between, and you discard that
>> job, the "Garbage collection" step will not be executed, so the garbages
>> will be left there.
>>
>> This is why we still recommend user to run offline cleanup every some
>> period; It is not perfert, but be good for most scenarios:
>> https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
>>
>> 2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
>>
>>> Woo... something new to me. Anybody knows?
>>>
>>> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com> wrote:
>>>
>>>> Kylin actually drops useless intermediate tables after cube building,
>>> but
>>>> I still see one "kylin_intermediate_cubename_searchdata" table for each
>>>> cube building in Hive. Are these tables still usefull for Kylin? I use
>>>> Kylin 1.5.2.1.
>>>>
>>>> Cheers,
>>>>
>>>> Jie
>>>>
>>
>>
>> --
>> Best regards,
>>
>> Shaofeng Shi
>>
>>
>
Re: kylin intermediate tables in Hive
Posted by ShaoFeng Shi <sh...@apache.org>.
BTW, are you using a view as lookup table?
2016-06-17 15:15 GMT+08:00 ShaoFeng Shi <sh...@apache.org>:
> This is common; If you have a job failed in between, and you discard that
> job, the "Garbage collection" step will not be executed, so the garbages
> will be left there.
>
> This is why we still recommend user to run offline cleanup every some
> period; It is not perfert, but be good for most scenarios:
> https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
>
> 2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
>
>> Woo... something new to me. Anybody knows?
>>
>> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com> wrote:
>>
>> > Kylin actually drops useless intermediate tables after cube building,
>> but
>> > I still see one "kylin_intermediate_cubename_searchdata" table for each
>> > cube building in Hive. Are these tables still usefull for Kylin? I use
>> > Kylin 1.5.2.1.
>> >
>> > Cheers,
>> >
>> > Jie
>> >
>>
>
>
>
> --
> Best regards,
>
> Shaofeng Shi
>
>
--
Best regards,
Shaofeng Shi
Re: kylin intermediate tables in Hive
Posted by ShaoFeng Shi <sh...@apache.org>.
This is common; If you have a job failed in between, and you discard that
job, the "Garbage collection" step will not be executed, so the garbages
will be left there.
This is why we still recommend user to run offline cleanup every some
period; It is not perfert, but be good for most scenarios:
https://kylin.apache.org/docs15/howto/howto_cleanup_storage.html
2016-06-17 15:00 GMT+08:00 Li Yang <li...@apache.org>:
> Woo... something new to me. Anybody knows?
>
> On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com> wrote:
>
> > Kylin actually drops useless intermediate tables after cube building, but
> > I still see one "kylin_intermediate_cubename_searchdata" table for each
> > cube building in Hive. Are these tables still usefull for Kylin? I use
> > Kylin 1.5.2.1.
> >
> > Cheers,
> >
> > Jie
> >
>
--
Best regards,
Shaofeng Shi
Re: kylin intermediate tables in Hive
Posted by Li Yang <li...@apache.org>.
Woo... something new to me. Anybody knows?
On Tue, Jun 14, 2016 at 6:57 PM, Jie Tao <ji...@gameforge.com> wrote:
> Kylin actually drops useless intermediate tables after cube building, but
> I still see one "kylin_intermediate_cubename_searchdata" table for each
> cube building in Hive. Are these tables still usefull for Kylin? I use
> Kylin 1.5.2.1.
>
> Cheers,
>
> Jie
>