hive-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mich Talebzadeh" <m...@peridale.co.uk>
Subject RE: Hive on Spark Engine versus Spark using Hive metastore
Date Wed, 03 Feb 2016 16:21:00 GMT
OK thanks. These are my new ENV settings based upon the availability of resources

 

export SPARK_EXECUTOR_CORES=12 ##, Number of cores for the workers (Default: 1).

export SPARK_EXECUTOR_MEMORY=5G ## , Memory per Worker (e.g. 1000M, 2G) (Default: 1G)

export SPARK_DRIVER_MEMORY=2G ## , Memory for Master (e.g. 1000M, 2G) (Default: 512 Mb)

 

These are the new runs after these settings:

 

Spark on Hive (3 consecutive runs)

 

 

spark-sql> select * from dummy where id in (1, 5, 100000);

1       0       0       63      rMLTDXxxqXOZnqYRJwInlGfGBTxNkAszBGEUGELqTSRnFjRGbi       
       1      xxxxxxxxxx

5       0       4       31      vDsFoYAOcitwrWNXCxPHzIIIxwKpTlrsVjFFKUDivytqJqOHGA       
       5      xxxxxxxxxx

100000  99      999     188     abQyrlxKzPTJliMqDpsfDTJUQzdNdfofUQhrKqXvRKwulZAoJe       
  100000      xxxxxxxxxx

Time taken: 47.987 seconds, Fetched 3 row(s)

 

Around 48 seconds

 

Hive on Spark 1.3.1

 

0: jdbc:hive2://rhes564:10010/default>  select * from dummy where id in (1, 5, 100000);

INFO  :

Query Hive on Spark job[2] stages:

INFO  : 2

INFO  :

Status: Running (Hive on Spark job[2])

INFO  : Job Progress Format

CurrentTime StageId_StageAttemptId: SucceededTasksCount(+RunningTasksCount-FailedTasksCount)/TotalTasksCount
[StageCost]

INFO  : 2016-02-03 16:20:50,315 Stage-2_0: 0(+18)/18

INFO  : 2016-02-03 16:20:53,369 Stage-2_0: 0(+18)/18

INFO  : 2016-02-03 16:20:56,478 Stage-2_0: 0(+18)/18

INFO  : 2016-02-03 16:20:58,530 Stage-2_0: 1(+17)/18

INFO  : 2016-02-03 16:21:01,570 Stage-2_0: 1(+17)/18

INFO  : 2016-02-03 16:21:04,680 Stage-2_0: 1(+17)/18

INFO  : 2016-02-03 16:21:07,767 Stage-2_0: 1(+17)/18

INFO  : 2016-02-03 16:21:10,877 Stage-2_0: 1(+17)/18

INFO  : 2016-02-03 16:21:13,941 Stage-2_0: 1(+17)/18

INFO  : 2016-02-03 16:21:17,019 Stage-2_0: 1(+17)/18

INFO  : 2016-02-03 16:21:20,090 Stage-2_0: 3(+15)/18

INFO  : 2016-02-03 16:21:21,138 Stage-2_0: 6(+12)/18

INFO  : 2016-02-03 16:21:22,145 Stage-2_0: 10(+8)/18

INFO  : 2016-02-03 16:21:23,150 Stage-2_0: 14(+4)/18

INFO  : 2016-02-03 16:21:24,154 Stage-2_0: 17(+1)/18

INFO  : 2016-02-03 16:21:26,161 Stage-2_0: 18/18 Finished

INFO  : Status: Finished successfully in 36.88 seconds

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

| dummy.id  | dummy.clustered  | dummy.scattered  | dummy.randomised  |                 dummy.random_string
                | dummy.small_vc  | dummy.padding  |

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

| 1         | 0                | 0                | 63                | rMLTDXxxqXOZnqYRJwInlGfGBTxNkAszBGEUGELqTSRnFjRGbi
 |          1      | xxxxxxxxxx     |

| 5         | 0                | 4                | 31                | vDsFoYAOcitwrWNXCxPHzIIIxwKpTlrsVjFFKUDivytqJqOHGA
 |          5      | xxxxxxxxxx     |

| 100000    | 99               | 999              | 188               | abQyrlxKzPTJliMqDpsfDTJUQzdNdfofUQhrKqXvRKwulZAoJe
 |     100000      | xxxxxxxxxx     |

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

3 rows selected (37.161 seconds)

 

Around 37 seconds

 

Interesting results

 

 

Dr Mich Talebzadeh

 

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

 

Sybase ASE 15 Gold Medal Award 2008

A Winning Strategy: Running the most Critical Financial Data on ASE 15

http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf

Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", ISBN 978-0-9563693-0-7.


co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 978-0-9759693-0-4

Publications due shortly:

Complex Event Processing in Heterogeneous Environments, ISBN: 978-0-9563693-3-8

Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume one out shortly

 

http://talebzadehmich.wordpress.com <http://talebzadehmich.wordpress.com/> 

 

NOTE: The information in this email is proprietary and confidential. This message is for the
designated recipient only, if you are not the intended recipient, you should destroy it immediately.
Any information in this message shall not be understood as given or endorsed by Peridale Technology
Ltd, its subsidiaries or their employees, unless expressly so stated. It is the responsibility
of the recipient to ensure that this email is virus free, therefore neither Peridale Technology
Ltd, its subsidiaries nor their employees accept any responsibility.

 

From: Xuefu Zhang [mailto:xzhang@cloudera.com] 
Sent: 03 February 2016 12:47
To: user@hive.apache.org
Subject: Re: Hive on Spark Engine versus Spark using Hive metastore

 

In YARN or standalone mode, you can set spark.executor.cores to utilize all cores on the node.
You can also set spark.executor.memory to allocate memory for Spark to use. Once you do this,
you may only have two executors to run your map tasks, but each core in each executor can
take up one task, increasing parallelism. With this, the eventually limit may come down to
the bandwidth of your disks in the cluster.

Having said that, a two-node cluster isn't really big enough to do performance benchmark.
Nevertheless, you still need to configure properly to make full use of the cluster.

--Xuefu

 

On Wed, Feb 3, 2016 at 1:25 AM, Mich Talebzadeh <mich@peridale.co.uk <mailto:mich@peridale.co.uk>
> wrote:

Hi Jeff,

 

I only have a two node cluster. Is there anyway one can simulate additional parallel runs
in such an environment thus having more than two maps?

 

thanks

 

Dr Mich Talebzadeh

 

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

 

Sybase ASE 15 Gold Medal Award 2008

A Winning Strategy: Running the most Critical Financial Data on ASE 15

http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf

Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", ISBN 978-0-9563693-0-7.


co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 978-0-9759693-0-4

Publications due shortly:

Complex Event Processing in Heterogeneous Environments, ISBN: 978-0-9563693-3-8

Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume one out shortly

 

http://talebzadehmich.wordpress.com <http://talebzadehmich.wordpress.com/> 

 

NOTE: The information in this email is proprietary and confidential. This message is for the
designated recipient only, if you are not the intended recipient, you should destroy it immediately.
Any information in this message shall not be understood as given or endorsed by Peridale Technology
Ltd, its subsidiaries or their employees, unless expressly so stated. It is the responsibility
of the recipient to ensure that this email is virus free, therefore neither Peridale Technology
Ltd, its subsidiaries nor their employees accept any responsibility.

 

From: Xuefu Zhang [mailto:xzhang@cloudera.com <mailto:xzhang@cloudera.com> ] 
Sent: 03 February 2016 02:39


To: user@hive.apache.org <mailto:user@hive.apache.org> 
Subject: Re: Hive on Spark Engine versus Spark using Hive metastore

 

Yes, regardless what spark mode you're running in, from Spark AM webui, you should be able
to see how many task are concurrently running. I'm a little surprised to see that your Hive
configuration only allows 2 map tasks to run in parallel. If your cluster has the capacity,
you should parallelize all the tasks to achieve optimal performance. Since I don't know your
Spark SQL configuration, I cannot tell how much parallelism you have over there. Thus, I'm
not sure if your comparison is valid.

--Xuefu

 

On Tue, Feb 2, 2016 at 5:08 PM, Mich Talebzadeh <mich@peridale.co.uk <mailto:mich@peridale.co.uk>
> wrote:

Hi Jeff,

 

In below

 

…. You should be able to see the resource usage in YARN resource manage URL.

 

Just to be clear we are talking about Port 8088/cluster?

 

Dr Mich Talebzadeh

 

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

 

Sybase ASE 15 Gold Medal Award 2008

A Winning Strategy: Running the most Critical Financial Data on ASE 15

http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf

Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", ISBN 978-0-9563693-0-7.


co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 978-0-9759693-0-4

Publications due shortly:

Complex Event Processing in Heterogeneous Environments, ISBN: 978-0-9563693-3-8

Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume one out shortly

 

http://talebzadehmich.wordpress.com <http://talebzadehmich.wordpress.com/> 

 

NOTE: The information in this email is proprietary and confidential. This message is for the
designated recipient only, if you are not the intended recipient, you should destroy it immediately.
Any information in this message shall not be understood as given or endorsed by Peridale Technology
Ltd, its subsidiaries or their employees, unless expressly so stated. It is the responsibility
of the recipient to ensure that this email is virus free, therefore neither Peridale Technology
Ltd, its subsidiaries nor their employees accept any responsibility.

 

From: Koert Kuipers [mailto:koert@tresata.com <mailto:koert@tresata.com> ] 
Sent: 03 February 2016 00:09

To: user@hive.apache.org <mailto:user@hive.apache.org> 
Subject: Re: Hive on Spark Engine versus Spark using Hive metastore

 

uuuhm with spark using Hive metastore you actually have a real programming environment and
you can write real functions, versus just being boxed into some version of sql and limited
udfs?

 

On Tue, Feb 2, 2016 at 6:46 PM, Xuefu Zhang <xzhang@cloudera.com <mailto:xzhang@cloudera.com>
> wrote:

When comparing the performance, you need to do it apple vs apple. In another thread, you mentioned
that Hive on Spark is much slower than Spark SQL. However, you configured Hive such that only
two tasks can run in parallel. However, you didn't provide information on how much Spark SQL
is utilizing. Thus, it's hard to tell whether it's just a configuration problem in your Hive
or Spark SQL is indeed faster. You should be able to see the resource usage in YARN resource
manage URL.

--Xuefu

 

On Tue, Feb 2, 2016 at 3:31 PM, Mich Talebzadeh <mich@peridale.co.uk <mailto:mich@peridale.co.uk>
> wrote:

Thanks Jeff.

 

Obviously Hive is much more feature rich compared to Spark. Having said that in certain areas
for example where the SQL feature is available in Spark, Spark seems to deliver faster.

 

This may be:

 

1.    Spark does both the optimisation and execution seamlessly

2.    Hive on Spark has to invoke YARN that adds another layer to the process

 

Now I did some simple tests on a 100Million rows ORC table available through Hive to both.

 

Spark 1.5.2 on Hive 1.2.1 Metastore

 

 

spark-sql> select * from dummy where id in (1, 5, 100000);

1       0       0       63      rMLTDXxxqXOZnqYRJwInlGfGBTxNkAszBGEUGELqTSRnFjRGbi       
       1      xxxxxxxxxx

5       0       4       31      vDsFoYAOcitwrWNXCxPHzIIIxwKpTlrsVjFFKUDivytqJqOHGA       
       5      xxxxxxxxxx

100000  99      999     188     abQyrlxKzPTJliMqDpsfDTJUQzdNdfofUQhrKqXvRKwulZAoJe       
  100000      xxxxxxxxxx

Time taken: 50.805 seconds, Fetched 3 row(s)

spark-sql> select * from dummy where id in (1, 5, 100000);

1       0       0       63      rMLTDXxxqXOZnqYRJwInlGfGBTxNkAszBGEUGELqTSRnFjRGbi       
       1      xxxxxxxxxx

5       0       4       31      vDsFoYAOcitwrWNXCxPHzIIIxwKpTlrsVjFFKUDivytqJqOHGA       
       5      xxxxxxxxxx

100000  99      999     188     abQyrlxKzPTJliMqDpsfDTJUQzdNdfofUQhrKqXvRKwulZAoJe       
  100000      xxxxxxxxxx

Time taken: 50.358 seconds, Fetched 3 row(s)

spark-sql> select * from dummy where id in (1, 5, 100000);

1       0       0       63      rMLTDXxxqXOZnqYRJwInlGfGBTxNkAszBGEUGELqTSRnFjRGbi       
       1      xxxxxxxxxx

5       0       4       31      vDsFoYAOcitwrWNXCxPHzIIIxwKpTlrsVjFFKUDivytqJqOHGA       
       5      xxxxxxxxxx

100000  99      999     188     abQyrlxKzPTJliMqDpsfDTJUQzdNdfofUQhrKqXvRKwulZAoJe       
  100000      xxxxxxxxxx

Time taken: 50.563 seconds, Fetched 3 row(s)

 

So three runs returning three rows just over 50 seconds

 

Hive 1.2.1 on spark 1.3.1 execution engine

 

0: jdbc:hive2://rhes564:10010/default> select * from dummy where id in (1, 5, 100000);

INFO  :

Query Hive on Spark job[4] stages:

INFO  : 4

INFO  :

Status: Running (Hive on Spark job[4])

INFO  : Status: Finished successfully in 82.49 seconds

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

| dummy.id <http://dummy.id>   | dummy.clustered  | dummy.scattered  | dummy.randomised
 |                 dummy.random_string                 | dummy.small_vc  | dummy.padding 
|

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

| 1         | 0                | 0                | 63                | rMLTDXxxqXOZnqYRJwInlGfGBTxNkAszBGEUGELqTSRnFjRGbi
 |          1      | xxxxxxxxxx     |

| 5         | 0                | 4                | 31                | vDsFoYAOcitwrWNXCxPHzIIIxwKpTlrsVjFFKUDivytqJqOHGA
 |          5      | xxxxxxxxxx     |

| 100000    | 99               | 999              | 188               | abQyrlxKzPTJliMqDpsfDTJUQzdNdfofUQhrKqXvRKwulZAoJe
 |     100000      | xxxxxxxxxx     |

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

3 rows selected (82.66 seconds)

0: jdbc:hive2://rhes564:10010/default> select * from dummy where id in (1, 5, 100000);

INFO  : Status: Finished successfully in 76.67 seconds

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

| dummy.id <http://dummy.id>   | dummy.clustered  | dummy.scattered  | dummy.randomised
 |                 dummy.random_string                 | dummy.small_vc  | dummy.padding 
|

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

| 1         | 0                | 0                | 63                | rMLTDXxxqXOZnqYRJwInlGfGBTxNkAszBGEUGELqTSRnFjRGbi
 |          1      | xxxxxxxxxx     |

| 5         | 0                | 4                | 31                | vDsFoYAOcitwrWNXCxPHzIIIxwKpTlrsVjFFKUDivytqJqOHGA
 |          5      | xxxxxxxxxx     |

| 100000    | 99               | 999              | 188               | abQyrlxKzPTJliMqDpsfDTJUQzdNdfofUQhrKqXvRKwulZAoJe
 |     100000      | xxxxxxxxxx     |

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

3 rows selected (76.835 seconds)

0: jdbc:hive2://rhes564:10010/default> select * from dummy where id in (1, 5, 100000);

INFO  : Status: Finished successfully in 80.54 seconds

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

| dummy.id <http://dummy.id>   | dummy.clustered  | dummy.scattered  | dummy.randomised
 |                 dummy.random_string                 | dummy.small_vc  | dummy.padding 
|

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

| 1         | 0                | 0                | 63                | rMLTDXxxqXOZnqYRJwInlGfGBTxNkAszBGEUGELqTSRnFjRGbi
 |          1      | xxxxxxxxxx     |

| 5         | 0                | 4                | 31                | vDsFoYAOcitwrWNXCxPHzIIIxwKpTlrsVjFFKUDivytqJqOHGA
 |          5      | xxxxxxxxxx     |

| 100000    | 99               | 999              | 188               | abQyrlxKzPTJliMqDpsfDTJUQzdNdfofUQhrKqXvRKwulZAoJe
 |     100000      | xxxxxxxxxx     |

+-----------+------------------+------------------+-------------------+-----------------------------------------------------+-----------------+----------------+--+

3 rows selected (80.718 seconds)

 

Three runs returning the same rows in 80 seconds. 

 

It is possible that My Spark engine with Hive is 1.3.1 which is out of date and that causes
this lag. 

 

There are certain queries that one cannot do with Spark. Besides it does not recognize CHAR
fields which is a pain.

 

spark-sql> CREATE TEMPORARY TABLE tmp AS

         > SELECT t.calendar_month_desc, c.channel_desc, SUM(s.amount_sold) AS TotalSales

         > FROM sales s, times t, channels c

         > WHERE s.time_id = t.time_id

         > AND   s.channel_id = c.channel_id

         > GROUP BY t.calendar_month_desc, c.channel_desc

         > ;

Error in query: Unhandled clauses: TEMPORARY 1, 2,2, 7

.

You are likely trying to use an unsupported Hive feature.";

 

 

 

 

 

Dr Mich Talebzadeh

 

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

 

Sybase ASE 15 Gold Medal Award 2008

A Winning Strategy: Running the most Critical Financial Data on ASE 15

http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf

Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", ISBN 978-0-9563693-0-7.


co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 978-0-9759693-0-4

Publications due shortly:

Complex Event Processing in Heterogeneous Environments, ISBN: 978-0-9563693-3-8

Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume one out shortly

 

http://talebzadehmich.wordpress.com <http://talebzadehmich.wordpress.com/> 

 

NOTE: The information in this email is proprietary and confidential. This message is for the
designated recipient only, if you are not the intended recipient, you should destroy it immediately.
Any information in this message shall not be understood as given or endorsed by Peridale Technology
Ltd, its subsidiaries or their employees, unless expressly so stated. It is the responsibility
of the recipient to ensure that this email is virus free, therefore neither Peridale Technology
Ltd, its subsidiaries nor their employees accept any responsibility.

 

From: Xuefu Zhang [mailto:xzhang@cloudera.com <mailto:xzhang@cloudera.com> ] 
Sent: 02 February 2016 23:12
To: user@hive.apache.org <mailto:user@hive.apache.org> 
Subject: Re: Hive on Spark Engine versus Spark using Hive metastore

 

I think the diff is not only about which does optimization but more on feature parity. Hive
on Spark offers all functional features that Hive offers and these features play out faster.
However, Spark SQL is far from offering this parity as far as I know.

 

On Tue, Feb 2, 2016 at 2:38 PM, Mich Talebzadeh <mich@peridale.co.uk <mailto:mich@peridale.co.uk>
> wrote:

Hi,

 

My understanding is that with Hive on Spark engine, one gets the Hive optimizer and Spark
query engine

 

With spark using Hive metastore, Spark does both the optimization and query engine. The only
value add is that one can access the underlying Hive tables from spark-sql etc

 

 

Is this assessment correct?

 

 

 

Thanks

 

Dr Mich Talebzadeh

 

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

 

Sybase ASE 15 Gold Medal Award 2008

A Winning Strategy: Running the most Critical Financial Data on ASE 15

http://login.sybase.com/files/Product_Overviews/ASE-Winning-Strategy-091908.pdf

Author of the books "A Practitioner’s Guide to Upgrading to Sybase ASE 15", ISBN 978-0-9563693-0-7.


co-author "Sybase Transact SQL Guidelines Best Practices", ISBN 978-0-9759693-0-4

Publications due shortly:

Complex Event Processing in Heterogeneous Environments, ISBN: 978-0-9563693-3-8

Oracle and Sybase, Concepts and Contrasts, ISBN: 978-0-9563693-1-4, volume one out shortly

 

http://talebzadehmich.wordpress.com <http://talebzadehmich.wordpress.com/> 

 

NOTE: The information in this email is proprietary and confidential. This message is for the
designated recipient only, if you are not the intended recipient, you should destroy it immediately.
Any information in this message shall not be understood as given or endorsed by Peridale Technology
Ltd, its subsidiaries or their employees, unless expressly so stated. It is the responsibility
of the recipient to ensure that this email is virus free, therefore neither Peridale Technology
Ltd, its subsidiaries nor their employees accept any responsibility.

 

 

 

 

 

 


Mime
View raw message