Re: Performance Question

[Benjamin Kim]

Todd,

I upgraded, deleted the table and recreated it again because it was 
unaccessible, and re-introduced the downed tablet server after clearing out all 
kudu directories.

The Spark Streaming job is repopulating again.

Thanks,
Ben


> On Jul 18, 2016, at 10:32 AM, Todd Lipcon  wrote:
> 
> On Mon, Jul 18, 2016 at 10:31 AM, Benjamin Kim  > wrote:
> Todd,
> 
> Thanks for the info. I was going to upgrade after the testing, but now, it 
> looks like I will have to do it earlier than expected.
> 
> I will do the upgrade, then resume.
> 
> OK, sounds good. The upgrade shouldn't invalidate any performance testing or 
> anything -- just fixes this important bug.
> 
> -Todd
> 
> 
>> On Jul 18, 2016, at 10:29 AM, Todd Lipcon > > wrote:
>> 
>> Hi Ben,
>> 
>> Any chance that you are running Kudu 0.9.0 instead of 0.9.1? There's a known 
>> serious bug in 0.9.0 which can cause this kind of corruption.
>> 
>> Assuming that you are running with replication count 3 this time, you should 
>> be able to move aside that tablet metadata file and start the server. It 
>> will recreate a new repaired replica automatically.
>> 
>> -Todd
>> 
>> On Mon, Jul 18, 2016 at 10:28 AM, Benjamin Kim > > wrote:
>> During my re-population of the Kudu table, I am getting this error trying to 
>> restart a tablet server after it went down. The job that populates this 
>> table has been running for over a week.
>> 
>> [libprotobuf ERROR google/protobuf/message_lite.cc:123] Can't parse message 
>> of type "kudu.tablet.TabletSuperBlockPB" because it is missing required 
>> fields: rowsets[2324].columns[15].block
>> F0718 17:01:26.783571   468 tablet_server_main.cc:55] Check failed: _s.ok() 
>> Bad status: IO error: Could not init Tablet Manager: Failed to open tablet 
>> metadata for tablet: 24637ee6f3e5440181ce3f20b1b298ba: Failed to load tablet 
>> metadata for tablet id 24637ee6f3e5440181ce3f20b1b298ba: Could not load 
>> tablet metadata from 
>> /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba: Unable to 
>> parse PB from path: 
>> /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba
>> *** Check failure stack trace: ***
>> @   0x7d794d  google::LogMessage::Fail()
>> @   0x7d984d  google::LogMessage::SendToLog()
>> @   0x7d7489  google::LogMessage::Flush()
>> @   0x7da2ef  google::LogMessageFatal::~LogMessageFatal()
>> @   0x78172b  (unknown)
>> @   0x344d41ed5d  (unknown)
>> @   0x7811d1  (unknown)
>> 
>> Does anyone know what this means?
>> 
>> Thanks,
>> Ben
>> 
>> 
>>> On Jul 11, 2016, at 10:47 AM, Todd Lipcon >> > wrote:
>>> 
>>> On Mon, Jul 11, 2016 at 10:40 AM, Benjamin Kim >> > wrote:
>>> Todd,
>>> 
>>> I had it at one replica. Do I have to recreate?
>>> 
>>> We don't currently have the ability to "accept data loss" on a tablet (or 
>>> set of tablets). If the machine is gone for good, then currently the only 
>>> easy way to recover is to recreate the table. If this sounds really 
>>> painful, though, maybe we can work up some kind of tool you could use to 
>>> just recreate the missing tablets (with those rows lost).
>>> 
>>> -Todd
>>> 
 On Jul 11, 2016, at 10:37 AM, Todd Lipcon > wrote:
 
 Hey Ben,
 
 Is the table that you're querying replicated? Or was it created with only 
 one replica per tablet?
 
 -Todd
 
 On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim > wrote:
 Over the weekend, a tablet server went down. It’s not coming back up. So, 
 I decommissioned it and removed it from the cluster. Then, I restarted 
 Kudu because I was getting a timeout  exception trying to do counts on the 
 table. Now, when I try again. I get the same error.
 
 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in stage 
 0.0 (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com 
 ): 
 com.stumbleupon.async.TimeoutException: Timed out after 3ms when 
 joining Deferred@712342716(state=PAUSED, result=Deferred@1765902299, 
 callback=passthrough -> scanner opened -> wakeup thread Executor task 
 launch worker-2, errback=openScanner errback -> passthrough -> wakeup 
 thread Executor task launch worker-2)
 at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
 at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
 at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
 at org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
 at 

Re: Performance Question

[Todd Lipcon]

On Mon, Jul 18, 2016 at 10:31 AM, Benjamin Kim  wrote:

> Todd,
>
> Thanks for the info. I was going to upgrade after the testing, but now, it
> looks like I will have to do it earlier than expected.
>
> I will do the upgrade, then resume.
>

OK, sounds good. The upgrade shouldn't invalidate any performance testing
or anything -- just fixes this important bug.

-Todd


> On Jul 18, 2016, at 10:29 AM, Todd Lipcon  wrote:
>
> Hi Ben,
>
> Any chance that you are running Kudu 0.9.0 instead of 0.9.1? There's a
> known serious bug in 0.9.0 which can cause this kind of corruption.
>
> Assuming that you are running with replication count 3 this time, you
> should be able to move aside that tablet metadata file and start the
> server. It will recreate a new repaired replica automatically.
>
> -Todd
>
> On Mon, Jul 18, 2016 at 10:28 AM, Benjamin Kim  wrote:
>
>> During my re-population of the Kudu table, I am getting this error trying
>> to restart a tablet server after it went down. The job that populates this
>> table has been running for over a week.
>>
>> [libprotobuf ERROR google/protobuf/message_lite.cc:123] Can't parse
>> message of type "kudu.tablet.TabletSuperBlockPB" because it is missing
>> required fields: rowsets[2324].columns[15].block
>> F0718 17:01:26.783571   468 tablet_server_main.cc:55] Check failed:
>> _s.ok() Bad status: IO error: Could not init Tablet Manager: Failed to open
>> tablet metadata for tablet: 24637ee6f3e5440181ce3f20b1b298ba: Failed to
>> load tablet metadata for tablet id 24637ee6f3e5440181ce3f20b1b298ba: Could
>> not load tablet metadata from
>> /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba: Unable
>> to parse PB from path:
>> /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba
>> *** Check failure stack trace: ***
>> @   0x7d794d  google::LogMessage::Fail()
>> @   0x7d984d  google::LogMessage::SendToLog()
>> @   0x7d7489  google::LogMessage::Flush()
>> @   0x7da2ef  google::LogMessageFatal::~LogMessageFatal()
>> @   0x78172b  (unknown)
>> @   0x344d41ed5d  (unknown)
>> @   0x7811d1  (unknown)
>>
>> Does anyone know what this means?
>>
>> Thanks,
>> Ben
>>
>>
>> On Jul 11, 2016, at 10:47 AM, Todd Lipcon  wrote:
>>
>> On Mon, Jul 11, 2016 at 10:40 AM, Benjamin Kim 
>> wrote:
>>
>>> Todd,
>>>
>>> I had it at one replica. Do I have to recreate?
>>>
>>
>> We don't currently have the ability to "accept data loss" on a tablet (or
>> set of tablets). If the machine is gone for good, then currently the only
>> easy way to recover is to recreate the table. If this sounds really
>> painful, though, maybe we can work up some kind of tool you could use to
>> just recreate the missing tablets (with those rows lost).
>>
>> -Todd
>>
>>>
>>> On Jul 11, 2016, at 10:37 AM, Todd Lipcon  wrote:
>>>
>>> Hey Ben,
>>>
>>> Is the table that you're querying replicated? Or was it created with
>>> only one replica per tablet?
>>>
>>> -Todd
>>>
>>> On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim  wrote:
>>>
 Over the weekend, a tablet server went down. It’s not coming back up.
 So, I decommissioned it and removed it from the cluster. Then, I restarted
 Kudu because I was getting a timeout  exception trying to do counts on the
 table. Now, when I try again. I get the same error.

 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in
 stage 0.0 (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com):
 com.stumbleupon.async.TimeoutException: Timed out after 3ms when
 joining Deferred@712342716(state=PAUSED, result=Deferred@1765902299,
 callback=passthrough -> scanner opened -> wakeup thread Executor task
 launch worker-2, errback=openScanner errback -> passthrough -> wakeup
 thread Executor task launch worker-2)
 at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
 at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
 at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
 at
 org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
 at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
 at
 org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:88)
 at
 org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:86)
 at
 org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
 at
 org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
 at
 org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
 at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
 at 

Re: Performance Question

[Benjamin Kim]

Todd,

Thanks for the info. I was going to upgrade after the testing, but now, it 
looks like I will have to do it earlier than expected.

I will do the upgrade, then resume.

Cheers,
Ben


> On Jul 18, 2016, at 10:29 AM, Todd Lipcon  wrote:
> 
> Hi Ben,
> 
> Any chance that you are running Kudu 0.9.0 instead of 0.9.1? There's a known 
> serious bug in 0.9.0 which can cause this kind of corruption.
> 
> Assuming that you are running with replication count 3 this time, you should 
> be able to move aside that tablet metadata file and start the server. It will 
> recreate a new repaired replica automatically.
> 
> -Todd
> 
> On Mon, Jul 18, 2016 at 10:28 AM, Benjamin Kim  > wrote:
> During my re-population of the Kudu table, I am getting this error trying to 
> restart a tablet server after it went down. The job that populates this table 
> has been running for over a week.
> 
> [libprotobuf ERROR google/protobuf/message_lite.cc:123] Can't parse message 
> of type "kudu.tablet.TabletSuperBlockPB" because it is missing required 
> fields: rowsets[2324].columns[15].block
> F0718 17:01:26.783571   468 tablet_server_main.cc:55] Check failed: _s.ok() 
> Bad status: IO error: Could not init Tablet Manager: Failed to open tablet 
> metadata for tablet: 24637ee6f3e5440181ce3f20b1b298ba: Failed to load tablet 
> metadata for tablet id 24637ee6f3e5440181ce3f20b1b298ba: Could not load 
> tablet metadata from 
> /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba: Unable to 
> parse PB from path: 
> /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba
> *** Check failure stack trace: ***
> @   0x7d794d  google::LogMessage::Fail()
> @   0x7d984d  google::LogMessage::SendToLog()
> @   0x7d7489  google::LogMessage::Flush()
> @   0x7da2ef  google::LogMessageFatal::~LogMessageFatal()
> @   0x78172b  (unknown)
> @   0x344d41ed5d  (unknown)
> @   0x7811d1  (unknown)
> 
> Does anyone know what this means?
> 
> Thanks,
> Ben
> 
> 
>> On Jul 11, 2016, at 10:47 AM, Todd Lipcon > > wrote:
>> 
>> On Mon, Jul 11, 2016 at 10:40 AM, Benjamin Kim > > wrote:
>> Todd,
>> 
>> I had it at one replica. Do I have to recreate?
>> 
>> We don't currently have the ability to "accept data loss" on a tablet (or 
>> set of tablets). If the machine is gone for good, then currently the only 
>> easy way to recover is to recreate the table. If this sounds really painful, 
>> though, maybe we can work up some kind of tool you could use to just 
>> recreate the missing tablets (with those rows lost).
>> 
>> -Todd
>> 
>>> On Jul 11, 2016, at 10:37 AM, Todd Lipcon >> > wrote:
>>> 
>>> Hey Ben,
>>> 
>>> Is the table that you're querying replicated? Or was it created with only 
>>> one replica per tablet?
>>> 
>>> -Todd
>>> 
>>> On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim >> > wrote:
>>> Over the weekend, a tablet server went down. It’s not coming back up. So, I 
>>> decommissioned it and removed it from the cluster. Then, I restarted Kudu 
>>> because I was getting a timeout  exception trying to do counts on the 
>>> table. Now, when I try again. I get the same error.
>>> 
>>> 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in stage 
>>> 0.0 (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com 
>>> ): 
>>> com.stumbleupon.async.TimeoutException: Timed out after 3ms when 
>>> joining Deferred@712342716(state=PAUSED, result=Deferred@1765902299, 
>>> callback=passthrough -> scanner opened -> wakeup thread Executor task 
>>> launch worker-2, errback=openScanner errback -> passthrough -> wakeup 
>>> thread Executor task launch worker-2)
>>> at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
>>> at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
>>> at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
>>> at org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
>>> at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
>>> at 
>>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:88)
>>> at 
>>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:86)
>>> at 
>>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>>> at 
>>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>>> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>>> at 

Re: Performance Question

[Todd Lipcon]

Hi Ben,

Any chance that you are running Kudu 0.9.0 instead of 0.9.1? There's a
known serious bug in 0.9.0 which can cause this kind of corruption.

Assuming that you are running with replication count 3 this time, you
should be able to move aside that tablet metadata file and start the
server. It will recreate a new repaired replica automatically.

-Todd

On Mon, Jul 18, 2016 at 10:28 AM, Benjamin Kim  wrote:

> During my re-population of the Kudu table, I am getting this error trying
> to restart a tablet server after it went down. The job that populates this
> table has been running for over a week.
>
> [libprotobuf ERROR google/protobuf/message_lite.cc:123] Can't parse
> message of type "kudu.tablet.TabletSuperBlockPB" because it is missing
> required fields: rowsets[2324].columns[15].block
> F0718 17:01:26.783571   468 tablet_server_main.cc:55] Check failed:
> _s.ok() Bad status: IO error: Could not init Tablet Manager: Failed to open
> tablet metadata for tablet: 24637ee6f3e5440181ce3f20b1b298ba: Failed to
> load tablet metadata for tablet id 24637ee6f3e5440181ce3f20b1b298ba: Could
> not load tablet metadata from
> /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba: Unable
> to parse PB from path:
> /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba
> *** Check failure stack trace: ***
> @   0x7d794d  google::LogMessage::Fail()
> @   0x7d984d  google::LogMessage::SendToLog()
> @   0x7d7489  google::LogMessage::Flush()
> @   0x7da2ef  google::LogMessageFatal::~LogMessageFatal()
> @   0x78172b  (unknown)
> @   0x344d41ed5d  (unknown)
> @   0x7811d1  (unknown)
>
> Does anyone know what this means?
>
> Thanks,
> Ben
>
>
> On Jul 11, 2016, at 10:47 AM, Todd Lipcon  wrote:
>
> On Mon, Jul 11, 2016 at 10:40 AM, Benjamin Kim  wrote:
>
>> Todd,
>>
>> I had it at one replica. Do I have to recreate?
>>
>
> We don't currently have the ability to "accept data loss" on a tablet (or
> set of tablets). If the machine is gone for good, then currently the only
> easy way to recover is to recreate the table. If this sounds really
> painful, though, maybe we can work up some kind of tool you could use to
> just recreate the missing tablets (with those rows lost).
>
> -Todd
>
>>
>> On Jul 11, 2016, at 10:37 AM, Todd Lipcon  wrote:
>>
>> Hey Ben,
>>
>> Is the table that you're querying replicated? Or was it created with only
>> one replica per tablet?
>>
>> -Todd
>>
>> On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim  wrote:
>>
>>> Over the weekend, a tablet server went down. It’s not coming back up.
>>> So, I decommissioned it and removed it from the cluster. Then, I restarted
>>> Kudu because I was getting a timeout  exception trying to do counts on the
>>> table. Now, when I try again. I get the same error.
>>>
>>> 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in
>>> stage 0.0 (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com):
>>> com.stumbleupon.async.TimeoutException: Timed out after 3ms when
>>> joining Deferred@712342716(state=PAUSED, result=Deferred@1765902299,
>>> callback=passthrough -> scanner opened -> wakeup thread Executor task
>>> launch worker-2, errback=openScanner errback -> passthrough -> wakeup
>>> thread Executor task launch worker-2)
>>> at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
>>> at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
>>> at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
>>> at org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
>>> at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
>>> at
>>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:88)
>>> at
>>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:86)
>>> at
>>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>>> at
>>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>>> at
>>> org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>>> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
>>> at
>>> org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>>> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
>>> at
>>> org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:73)
>>> at
>>> org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)
>>> at org.apache.spark.scheduler.Task.run(Task.scala:89)
>>> at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:214)
>>> at
>>> 

Re: Performance Question

[Benjamin Kim]

During my re-population of the Kudu table, I am getting this error trying to 
restart a tablet server after it went down. The job that populates this table 
has been running for over a week.

[libprotobuf ERROR google/protobuf/message_lite.cc:123] Can't parse message of 
type "kudu.tablet.TabletSuperBlockPB" because it is missing required fields: 
rowsets[2324].columns[15].block
F0718 17:01:26.783571   468 tablet_server_main.cc:55] Check failed: _s.ok() Bad 
status: IO error: Could not init Tablet Manager: Failed to open tablet metadata 
for tablet: 24637ee6f3e5440181ce3f20b1b298ba: Failed to load tablet metadata 
for tablet id 24637ee6f3e5440181ce3f20b1b298ba: Could not load tablet metadata 
from /mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba: Unable 
to parse PB from path: 
/mnt/data1/kudu/data/tablet-meta/24637ee6f3e5440181ce3f20b1b298ba
*** Check failure stack trace: ***
@   0x7d794d  google::LogMessage::Fail()
@   0x7d984d  google::LogMessage::SendToLog()
@   0x7d7489  google::LogMessage::Flush()
@   0x7da2ef  google::LogMessageFatal::~LogMessageFatal()
@   0x78172b  (unknown)
@   0x344d41ed5d  (unknown)
@   0x7811d1  (unknown)

Does anyone know what this means?

Thanks,
Ben


> On Jul 11, 2016, at 10:47 AM, Todd Lipcon  wrote:
> 
> On Mon, Jul 11, 2016 at 10:40 AM, Benjamin Kim  > wrote:
> Todd,
> 
> I had it at one replica. Do I have to recreate?
> 
> We don't currently have the ability to "accept data loss" on a tablet (or set 
> of tablets). If the machine is gone for good, then currently the only easy 
> way to recover is to recreate the table. If this sounds really painful, 
> though, maybe we can work up some kind of tool you could use to just recreate 
> the missing tablets (with those rows lost).
> 
> -Todd
> 
>> On Jul 11, 2016, at 10:37 AM, Todd Lipcon > > wrote:
>> 
>> Hey Ben,
>> 
>> Is the table that you're querying replicated? Or was it created with only 
>> one replica per tablet?
>> 
>> -Todd
>> 
>> On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim > > wrote:
>> Over the weekend, a tablet server went down. It’s not coming back up. So, I 
>> decommissioned it and removed it from the cluster. Then, I restarted Kudu 
>> because I was getting a timeout  exception trying to do counts on the table. 
>> Now, when I try again. I get the same error.
>> 
>> 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in stage 
>> 0.0 (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com 
>> ): 
>> com.stumbleupon.async.TimeoutException: Timed out after 3ms when joining 
>> Deferred@712342716(state=PAUSED, result=Deferred@1765902299, 
>> callback=passthrough -> scanner opened -> wakeup thread Executor task launch 
>> worker-2, errback=openScanner errback -> passthrough -> wakeup thread 
>> Executor task launch worker-2)
>> at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
>> at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
>> at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
>> at org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
>> at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
>> at 
>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:88)
>> at 
>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:86)
>> at 
>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>> at 
>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
>> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
>> at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:73)
>> at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)
>> at org.apache.spark.scheduler.Task.run(Task.scala:89)
>> at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:214)
>> at 
>> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
>> at 
>> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
>> at java.lang.Thread.run(Thread.java:745)
>> 
>> Does anyone know how to recover from this?
>> 
>> Thanks,
>> Benjamin Kim
>> Data Solutions Architect
>> 
>> [a•mo•bee] (n.) the company defining digital marketing.
>> 

Re: Performance Question

[Benjamin Kim]

Todd,

It’s no problem to start over again. But, a tool like that would be helpful. 
Gaps in data can be accommodated for by just back filling.

Thanks,
Ben

> On Jul 11, 2016, at 10:47 AM, Todd Lipcon  wrote:
> 
> On Mon, Jul 11, 2016 at 10:40 AM, Benjamin Kim  > wrote:
> Todd,
> 
> I had it at one replica. Do I have to recreate?
> 
> We don't currently have the ability to "accept data loss" on a tablet (or set 
> of tablets). If the machine is gone for good, then currently the only easy 
> way to recover is to recreate the table. If this sounds really painful, 
> though, maybe we can work up some kind of tool you could use to just recreate 
> the missing tablets (with those rows lost).
> 
> -Todd
> 
>> On Jul 11, 2016, at 10:37 AM, Todd Lipcon > > wrote:
>> 
>> Hey Ben,
>> 
>> Is the table that you're querying replicated? Or was it created with only 
>> one replica per tablet?
>> 
>> -Todd
>> 
>> On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim > > wrote:
>> Over the weekend, a tablet server went down. It’s not coming back up. So, I 
>> decommissioned it and removed it from the cluster. Then, I restarted Kudu 
>> because I was getting a timeout  exception trying to do counts on the table. 
>> Now, when I try again. I get the same error.
>> 
>> 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in stage 
>> 0.0 (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com 
>> ): 
>> com.stumbleupon.async.TimeoutException: Timed out after 3ms when joining 
>> Deferred@712342716(state=PAUSED, result=Deferred@1765902299, 
>> callback=passthrough -> scanner opened -> wakeup thread Executor task launch 
>> worker-2, errback=openScanner errback -> passthrough -> wakeup thread 
>> Executor task launch worker-2)
>> at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
>> at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
>> at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
>> at org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
>> at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
>> at 
>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:88)
>> at 
>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:86)
>> at 
>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>> at 
>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
>> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
>> at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:73)
>> at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)
>> at org.apache.spark.scheduler.Task.run(Task.scala:89)
>> at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:214)
>> at 
>> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
>> at 
>> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
>> at java.lang.Thread.run(Thread.java:745)
>> 
>> Does anyone know how to recover from this?
>> 
>> Thanks,
>> Benjamin Kim
>> Data Solutions Architect
>> 
>> [a•mo•bee] (n.) the company defining digital marketing.
>> 
>> Mobile: +1 818 635 2900 
>> 3250 Ocean Park Blvd, Suite 200  |  Santa Monica, CA 90405  |  
>> www.amobee.com 
>>> On Jul 6, 2016, at 9:46 AM, Dan Burkert >> > wrote:
>>> 
>>> 
>>> 
>>> On Wed, Jul 6, 2016 at 7:05 AM, Benjamin Kim >> > wrote:
>>> Over the weekend, the row count is up to <500M. I will give it another few 
>>> days to get to 1B rows. I still get consistent times ~15s for doing row 
>>> counts despite the amount of data growing.
>>> 
>>> On another note, I got a solicitation email from SnappyData to evaluate 
>>> their product. They claim to be the “Spark Data Store” with tight 
>>> integration with Spark executors. It claims to be an OLTP and OLAP system 
>>> with being an in-memory data store first then to disk. After going to 
>>> several Spark events, it would seem that this is the new “hot” area for 
>>> vendors. They all (MemSQL, Redis, Aerospike, Datastax, etc.) claim to be 
>>> the best "Spark Data Store”. I’m wondering if Kudu will become this too? 

Re: Performance Question

[Todd Lipcon]

On Mon, Jul 11, 2016 at 10:40 AM, Benjamin Kim  wrote:

> Todd,
>
> I had it at one replica. Do I have to recreate?
>

We don't currently have the ability to "accept data loss" on a tablet (or
set of tablets). If the machine is gone for good, then currently the only
easy way to recover is to recreate the table. If this sounds really
painful, though, maybe we can work up some kind of tool you could use to
just recreate the missing tablets (with those rows lost).

-Todd

>
> On Jul 11, 2016, at 10:37 AM, Todd Lipcon  wrote:
>
> Hey Ben,
>
> Is the table that you're querying replicated? Or was it created with only
> one replica per tablet?
>
> -Todd
>
> On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim  wrote:
>
>> Over the weekend, a tablet server went down. It’s not coming back up. So,
>> I decommissioned it and removed it from the cluster. Then, I restarted Kudu
>> because I was getting a timeout  exception trying to do counts on the
>> table. Now, when I try again. I get the same error.
>>
>> 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in stage
>> 0.0 (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com):
>> com.stumbleupon.async.TimeoutException: Timed out after 3ms when
>> joining Deferred@712342716(state=PAUSED, result=Deferred@1765902299,
>> callback=passthrough -> scanner opened -> wakeup thread Executor task
>> launch worker-2, errback=openScanner errback -> passthrough -> wakeup
>> thread Executor task launch worker-2)
>> at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
>> at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
>> at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
>> at org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
>> at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
>> at
>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:88)
>> at
>> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:86)
>> at
>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>> at
>> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
>> at
>> org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
>> at
>> org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
>> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
>> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
>> at
>> org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:73)
>> at
>> org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)
>> at org.apache.spark.scheduler.Task.run(Task.scala:89)
>> at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:214)
>> at
>> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
>> at
>> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
>> at java.lang.Thread.run(Thread.java:745)
>>
>> Does anyone know how to recover from this?
>>
>> Thanks,
>> *Benjamin Kim*
>> *Data Solutions Architect*
>>
>> [a•mo•bee] *(n.)* the company defining digital marketing.
>>
>> *Mobile: +1 818 635 2900 <%2B1%20818%20635%202900>*
>> 3250 Ocean Park Blvd, Suite 200  |  Santa Monica, CA 90405  |
>> www.amobee.com
>>
>> On Jul 6, 2016, at 9:46 AM, Dan Burkert  wrote:
>>
>>
>>
>> On Wed, Jul 6, 2016 at 7:05 AM, Benjamin Kim  wrote:
>>
>>> Over the weekend, the row count is up to <500M. I will give it another
>>> few days to get to 1B rows. I still get consistent times ~15s for doing row
>>> counts despite the amount of data growing.
>>>
>>> On another note, I got a solicitation email from SnappyData to evaluate
>>> their product. They claim to be the “Spark Data Store” with tight
>>> integration with Spark executors. It claims to be an OLTP and OLAP system
>>> with being an in-memory data store first then to disk. After going to
>>> several Spark events, it would seem that this is the new “hot” area for
>>> vendors. They all (MemSQL, Redis, Aerospike, Datastax, etc.) claim to be
>>> the best "Spark Data Store”. I’m wondering if Kudu will become this too?
>>> With the performance I’ve seen so far, it would seem that it can be a
>>> contender. All that is needed is a hardened Spark connector package, I
>>> would think. The next evaluation I will be conducting is to see if
>>> SnappyData’s claims are valid by doing my own tests.
>>>
>>
>> It's hard to compare Kudu against any other data store without a lot of
>> analysis and thorough benchmarking, but it is certainly a goal of Kudu to
>> be a great platform for ingesting and analyzing data through Spark.  Up
>> till 

Re: Performance Question

[Benjamin Kim]

Todd,

I had it at one replica. Do I have to recreate?

Thanks,
Ben


> On Jul 11, 2016, at 10:37 AM, Todd Lipcon  wrote:
> 
> Hey Ben,
> 
> Is the table that you're querying replicated? Or was it created with only one 
> replica per tablet?
> 
> -Todd
> 
> On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim  > wrote:
> Over the weekend, a tablet server went down. It’s not coming back up. So, I 
> decommissioned it and removed it from the cluster. Then, I restarted Kudu 
> because I was getting a timeout  exception trying to do counts on the table. 
> Now, when I try again. I get the same error.
> 
> 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in stage 0.0 
> (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com 
> ): 
> com.stumbleupon.async.TimeoutException: Timed out after 3ms when joining 
> Deferred@712342716(state=PAUSED, result=Deferred@1765902299, 
> callback=passthrough -> scanner opened -> wakeup thread Executor task launch 
> worker-2, errback=openScanner errback -> passthrough -> wakeup thread 
> Executor task launch worker-2)
> at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
> at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
> at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
> at org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
> at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
> at 
> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:88)
> at 
> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:86)
> at 
> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
> at 
> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
> at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:73)
> at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)
> at org.apache.spark.scheduler.Task.run(Task.scala:89)
> at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:214)
> at 
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at 
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> 
> Does anyone know how to recover from this?
> 
> Thanks,
> Benjamin Kim
> Data Solutions Architect
> 
> [a•mo•bee] (n.) the company defining digital marketing.
> 
> Mobile: +1 818 635 2900 
> 3250 Ocean Park Blvd, Suite 200  |  Santa Monica, CA 90405  |  www.amobee.com 
> 
>> On Jul 6, 2016, at 9:46 AM, Dan Burkert > > wrote:
>> 
>> 
>> 
>> On Wed, Jul 6, 2016 at 7:05 AM, Benjamin Kim > > wrote:
>> Over the weekend, the row count is up to <500M. I will give it another few 
>> days to get to 1B rows. I still get consistent times ~15s for doing row 
>> counts despite the amount of data growing.
>> 
>> On another note, I got a solicitation email from SnappyData to evaluate 
>> their product. They claim to be the “Spark Data Store” with tight 
>> integration with Spark executors. It claims to be an OLTP and OLAP system 
>> with being an in-memory data store first then to disk. After going to 
>> several Spark events, it would seem that this is the new “hot” area for 
>> vendors. They all (MemSQL, Redis, Aerospike, Datastax, etc.) claim to be the 
>> best "Spark Data Store”. I’m wondering if Kudu will become this too? With 
>> the performance I’ve seen so far, it would seem that it can be a contender. 
>> All that is needed is a hardened Spark connector package, I would think. The 
>> next evaluation I will be conducting is to see if SnappyData’s claims are 
>> valid by doing my own tests.
>> 
>> It's hard to compare Kudu against any other data store without a lot of 
>> analysis and thorough benchmarking, but it is certainly a goal of Kudu to be 
>> a great platform for ingesting and analyzing data through Spark.  Up till 
>> this point most of the Spark work has been community driven, but more 
>> thorough integration testing of the Spark connector is going to be a focus 
>> going forward.
>> 
>> - Dan
>> 
>>  
>> Cheers,
>> Ben
>> 
>> 
>> 
>>> On Jun 15, 2016, at 12:47 AM, Todd Lipcon >> 

Re: Performance Question

[Todd Lipcon]

Hey Ben,

Is the table that you're querying replicated? Or was it created with only
one replica per tablet?

-Todd

On Mon, Jul 11, 2016 at 10:35 AM, Benjamin Kim  wrote:

> Over the weekend, a tablet server went down. It’s not coming back up. So,
> I decommissioned it and removed it from the cluster. Then, I restarted Kudu
> because I was getting a timeout  exception trying to do counts on the
> table. Now, when I try again. I get the same error.
>
> 16/07/11 17:32:36 WARN scheduler.TaskSetManager: Lost task 468.3 in stage
> 0.0 (TID 603, prod-dc1-datanode167.pdc1i.gradientx.com):
> com.stumbleupon.async.TimeoutException: Timed out after 3ms when
> joining Deferred@712342716(state=PAUSED, result=Deferred@1765902299,
> callback=passthrough -> scanner opened -> wakeup thread Executor task
> launch worker-2, errback=openScanner errback -> passthrough -> wakeup
> thread Executor task launch worker-2)
> at com.stumbleupon.async.Deferred.doJoin(Deferred.java:1177)
> at com.stumbleupon.async.Deferred.join(Deferred.java:1045)
> at org.kududb.client.KuduScanner.nextRows(KuduScanner.java:57)
> at org.kududb.spark.kudu.RowResultIteratorScala.hasNext(KuduRDD.scala:99)
> at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
> at
> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:88)
> at
> org.apache.spark.sql.execution.aggregate.TungstenAggregate$$anonfun$doExecute$1$$anonfun$2.apply(TungstenAggregate.scala:86)
> at
> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
> at
> org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
> at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
> at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
> at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
> at
> org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:73)
> at
> org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)
> at org.apache.spark.scheduler.Task.run(Task.scala:89)
> at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:214)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
>
> Does anyone know how to recover from this?
>
> Thanks,
> *Benjamin Kim*
> *Data Solutions Architect*
>
> [a•mo•bee] *(n.)* the company defining digital marketing.
>
> *Mobile: +1 818 635 2900 <%2B1%20818%20635%202900>*
> 3250 Ocean Park Blvd, Suite 200  |  Santa Monica, CA 90405  |
> www.amobee.com
>
> On Jul 6, 2016, at 9:46 AM, Dan Burkert  wrote:
>
>
>
> On Wed, Jul 6, 2016 at 7:05 AM, Benjamin Kim  wrote:
>
>> Over the weekend, the row count is up to <500M. I will give it another
>> few days to get to 1B rows. I still get consistent times ~15s for doing row
>> counts despite the amount of data growing.
>>
>> On another note, I got a solicitation email from SnappyData to evaluate
>> their product. They claim to be the “Spark Data Store” with tight
>> integration with Spark executors. It claims to be an OLTP and OLAP system
>> with being an in-memory data store first then to disk. After going to
>> several Spark events, it would seem that this is the new “hot” area for
>> vendors. They all (MemSQL, Redis, Aerospike, Datastax, etc.) claim to be
>> the best "Spark Data Store”. I’m wondering if Kudu will become this too?
>> With the performance I’ve seen so far, it would seem that it can be a
>> contender. All that is needed is a hardened Spark connector package, I
>> would think. The next evaluation I will be conducting is to see if
>> SnappyData’s claims are valid by doing my own tests.
>>
>
> It's hard to compare Kudu against any other data store without a lot of
> analysis and thorough benchmarking, but it is certainly a goal of Kudu to
> be a great platform for ingesting and analyzing data through Spark.  Up
> till this point most of the Spark work has been community driven, but more
> thorough integration testing of the Spark connector is going to be a focus
> going forward.
>
> - Dan
>
>
>
>> Cheers,
>> Ben
>>
>>
>>
>> On Jun 15, 2016, at 12:47 AM, Todd Lipcon  wrote:
>>
>> Hi Benjamin,
>>
>> What workload are you using for benchmarks? Using spark or something more
>> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and
>> some queries
>>
>> Todd
>>
>> Todd
>> On Jun 15, 2016 8:10 AM, "Benjamin Kim"  wrote:
>>
>>> Hi Todd,
>>>
>>> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am
>>> impressed. 

Re: Performance Question

[Benjamin Kim]

Dan,

This is good to hear as we are heavily invested in Spark as are many of our 
competitors in the AdTech/Telecom world. It would be nice to have Kudu be on 
par with the other data store technologies in terms of Spark usability, so as 
to not choose one based on “who provides it now in production”, as management 
tends to say.

Cheers,
Ben

> On Jul 6, 2016, at 9:46 AM, Dan Burkert  wrote:
> 
> 
> 
> On Wed, Jul 6, 2016 at 7:05 AM, Benjamin Kim  > wrote:
> Over the weekend, the row count is up to <500M. I will give it another few 
> days to get to 1B rows. I still get consistent times ~15s for doing row 
> counts despite the amount of data growing.
> 
> On another note, I got a solicitation email from SnappyData to evaluate their 
> product. They claim to be the “Spark Data Store” with tight integration with 
> Spark executors. It claims to be an OLTP and OLAP system with being an 
> in-memory data store first then to disk. After going to several Spark events, 
> it would seem that this is the new “hot” area for vendors. They all (MemSQL, 
> Redis, Aerospike, Datastax, etc.) claim to be the best "Spark Data Store”. 
> I’m wondering if Kudu will become this too? With the performance I’ve seen so 
> far, it would seem that it can be a contender. All that is needed is a 
> hardened Spark connector package, I would think. The next evaluation I will 
> be conducting is to see if SnappyData’s claims are valid by doing my own 
> tests.
> 
> It's hard to compare Kudu against any other data store without a lot of 
> analysis and thorough benchmarking, but it is certainly a goal of Kudu to be 
> a great platform for ingesting and analyzing data through Spark.  Up till 
> this point most of the Spark work has been community driven, but more 
> thorough integration testing of the Spark connector is going to be a focus 
> going forward.
> 
> - Dan
> 
>  
> Cheers,
> Ben
> 
> 
> 
>> On Jun 15, 2016, at 12:47 AM, Todd Lipcon > > wrote:
>> 
>> Hi Benjamin,
>> 
>> What workload are you using for benchmarks? Using spark or something more 
>> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and 
>> some queries
>> 
>> Todd
>> 
>> Todd
>> 
>> On Jun 15, 2016 8:10 AM, "Benjamin Kim" > > wrote:
>> Hi Todd,
>> 
>> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am impressed. 
>> Compared to HBase, read and write performance are better. Write performance 
>> has the greatest improvement (> 4x), while read is > 1.5x. Albeit, these are 
>> only preliminary tests. Do you know of a way to really do some conclusive 
>> tests? I want to see if I can match your results on my 50 node cluster.
>> 
>> Thanks,
>> Ben
>> 
>>> On May 30, 2016, at 10:33 AM, Todd Lipcon >> > wrote:
>>> 
>>> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim >> > wrote:
>>> Todd,
>>> 
>>> It sounds like Kudu can possibly top or match those numbers put out by 
>>> Aerospike. Do you have any performance statistics published or any 
>>> instructions as to measure them myself as good way to test? In addition, 
>>> this will be a test using Spark, so should I wait for Kudu version 0.9.0 
>>> where support will be built in?
>>> 
>>> We don't have a lot of benchmarks published yet, especially on the write 
>>> side. I've found that thorough cross-system benchmarks are very difficult 
>>> to do fairly and accurately, and often times users end up misguided if they 
>>> pay too much attention to them :) So, given a finite number of developers 
>>> working on Kudu, I think we've tended to spend more time on the project 
>>> itself and less time focusing on "competition". I'm sure there are use 
>>> cases where Kudu will beat out Aerospike, and probably use cases where 
>>> Aerospike will beat Kudu as well.
>>> 
>>> From my perspective, it would be great if you can share some details of 
>>> your workload, especially if there are some areas you're finding Kudu 
>>> lacking. Maybe we can spot some easy code changes we could make to improve 
>>> performance, or suggest a tuning variable you could change.
>>> 
>>> -Todd
>>> 
>>> 
 On May 27, 2016, at 9:19 PM, Todd Lipcon > wrote:
 
 On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim > wrote:
 Hi Mike,
 
 First of all, thanks for the link. It looks like an interesting read. I 
 checked that Aerospike is currently at version 3.8.2.3, and in the 
 article, they are evaluating version 3.5.4. The main thing that impressed 
 me was their claim that they can beat Cassandra and HBase by 8x for 
 writing and 25x for reading. Their big claim to fame is that Aerospike can 
 write 1M records per second with 

Re: Performance Question

[Dan Burkert]

On Wed, Jul 6, 2016 at 7:05 AM, Benjamin Kim  wrote:

> Over the weekend, the row count is up to <500M. I will give it another few
> days to get to 1B rows. I still get consistent times ~15s for doing row
> counts despite the amount of data growing.
>
> On another note, I got a solicitation email from SnappyData to evaluate
> their product. They claim to be the “Spark Data Store” with tight
> integration with Spark executors. It claims to be an OLTP and OLAP system
> with being an in-memory data store first then to disk. After going to
> several Spark events, it would seem that this is the new “hot” area for
> vendors. They all (MemSQL, Redis, Aerospike, Datastax, etc.) claim to be
> the best "Spark Data Store”. I’m wondering if Kudu will become this too?
> With the performance I’ve seen so far, it would seem that it can be a
> contender. All that is needed is a hardened Spark connector package, I
> would think. The next evaluation I will be conducting is to see if
> SnappyData’s claims are valid by doing my own tests.
>

It's hard to compare Kudu against any other data store without a lot of
analysis and thorough benchmarking, but it is certainly a goal of Kudu to
be a great platform for ingesting and analyzing data through Spark.  Up
till this point most of the Spark work has been community driven, but more
thorough integration testing of the Spark connector is going to be a focus
going forward.

- Dan



> Cheers,
> Ben
>
>
>
> On Jun 15, 2016, at 12:47 AM, Todd Lipcon  wrote:
>
> Hi Benjamin,
>
> What workload are you using for benchmarks? Using spark or something more
> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and
> some queries
>
> Todd
>
> Todd
> On Jun 15, 2016 8:10 AM, "Benjamin Kim"  wrote:
>
>> Hi Todd,
>>
>> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am
>> impressed. Compared to HBase, read and write performance are better. Write
>> performance has the greatest improvement (> 4x), while read is > 1.5x.
>> Albeit, these are only preliminary tests. Do you know of a way to really do
>> some conclusive tests? I want to see if I can match your results on my 50
>> node cluster.
>>
>> Thanks,
>> Ben
>>
>> On May 30, 2016, at 10:33 AM, Todd Lipcon  wrote:
>>
>> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim  wrote:
>>
>>> Todd,
>>>
>>> It sounds like Kudu can possibly top or match those numbers put out by
>>> Aerospike. Do you have any performance statistics published or any
>>> instructions as to measure them myself as good way to test? In addition,
>>> this will be a test using Spark, so should I wait for Kudu version 0.9.0
>>> where support will be built in?
>>>
>>
>> We don't have a lot of benchmarks published yet, especially on the write
>> side. I've found that thorough cross-system benchmarks are very difficult
>> to do fairly and accurately, and often times users end up misguided if they
>> pay too much attention to them :) So, given a finite number of developers
>> working on Kudu, I think we've tended to spend more time on the project
>> itself and less time focusing on "competition". I'm sure there are use
>> cases where Kudu will beat out Aerospike, and probably use cases where
>> Aerospike will beat Kudu as well.
>>
>> From my perspective, it would be great if you can share some details of
>> your workload, especially if there are some areas you're finding Kudu
>> lacking. Maybe we can spot some easy code changes we could make to improve
>> performance, or suggest a tuning variable you could change.
>>
>> -Todd
>>
>>
>>> On May 27, 2016, at 9:19 PM, Todd Lipcon  wrote:
>>>
>>> On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim 
>>> wrote:
>>>
 Hi Mike,

 First of all, thanks for the link. It looks like an interesting read. I
 checked that Aerospike is currently at version 3.8.2.3, and in the article,
 they are evaluating version 3.5.4. The main thing that impressed me was
 their claim that they can beat Cassandra and HBase by 8x for writing and
 25x for reading. Their big claim to fame is that Aerospike can write 1M
 records per second with only 50 nodes. I wanted to see if this is real.

>>>
>>> 1M records per second on 50 nodes is pretty doable by Kudu as well,
>>> depending on the size of your records and the insertion order. I've been
>>> playing with a ~70 node cluster recently and seen 1M+ writes/second
>>> sustained, and bursting above 4M. These are 1KB rows with 11 columns, and
>>> with pretty old HDD-only nodes. I think newer flash-based nodes could do
>>> better.
>>>
>>>

 To answer your questions, we have a DMP with user profiles with many
 attributes. We create segmentation information off of these attributes to
 classify them. Then, we can target advertising appropriately for our sales
 department. Much of the data processing is for applying 

Re: Performance Question

[Dan Burkert]

On Mon, Jul 4, 2016 at 2:46 AM, 袁康（梓悠） <yuankang...@alibaba-inc.com> wrote:

> How can I delete data in kudu table wiht spark  (not delete the table at
> all)?
>

We do not currently have a way to delete a Kudu table through the spark
connector, but you should be able to instantiate a Kudu client and delete
the table that way.  We have discussed making one of the spark write modes
do a truncate operation, but nothing has been implemented.

 - Dan


> --
> 发件人：Todd Lipcon <t...@cloudera.com>
> 发送时间：2016年7月2日(星期六) 02:44
> 收件人：user <user@kudu.incubator.apache.org>
> 主 题：Re: Performance Question
>
> On Thu, Jun 30, 2016 at 5:39 PM, Benjamin Kim <bbuil...@gmail.com> wrote:
> Hi Todd,
>
> I changed the key to be what you suggested, and I can’t tell the
> difference since it was already fast. But, I did get more numbers.
>
> Yea, you won't see a substantial difference until you're inserting
> billions of rows, etc, and the keys and/or bloom filters no longer fit in
> cache.
>
>
> > 104M rows in Kudu table
> - read: 8s
> - count: 16s
> - aggregate: 9s
>
> The time to read took much longer from 0.2s to 8s, counts were the same
> 16s, and aggregate queries look longer from 6s to 9s.
>
> I’m still impressed.
>
> We aim to please ;-) If you have any interest in writing up these
> experiments as a blog post, would be cool to post them for others to learn
> from.
>
> -Todd
>
> On Jun 15, 2016, at 12:47 AM, Todd Lipcon <t...@cloudera.com> wrote:
>
> Hi Benjamin,
>
> What workload are you using for benchmarks? Using spark or something more
> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and
> some queries
>
> Todd
>
> Todd
> On Jun 15, 2016 8:10 AM, "Benjamin Kim" <bbuil...@gmail.com> wrote:
> Hi Todd,
>
> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am
> impressed. Compared to HBase, read and write performance are better. Write
> performance has the greatest improvement (> 4x), while read is > 1.5x.
> Albeit, these are only preliminary tests. Do you know of a way to really do
> some conclusive tests? I want to see if I can match your results on my 50
> node cluster.
>
> Thanks,
> Ben
>
> On May 30, 2016, at 10:33 AM, Todd Lipcon <t...@cloudera.com> wrote:
>
> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim <bbuil...@gmail.com> wrote:
> Todd,
>
> It sounds like Kudu can possibly top or match those numbers put out by
> Aerospike. Do you have any performance statistics published or any
> instructions as to measure them myself as good way to test? In addition,
> this will be a test using Spark, so should I wait for Kudu version 0.9.0
> where support will be built in?
>
> We don't have a lot of benchmarks published yet, especially on the write
> side. I've found that thorough cross-system benchmarks are very difficult
> to do fairly and accurately, and often times users end up misguided if they
> pay too much attention to them :) So, given a finite number of developers
> working on Kudu, I think we've tended to spend more time on the project
> itself and less time focusing on "competition". I'm sure there are use
> cases where Kudu will beat out Aerospike, and probably use cases where
> Aerospike will beat Kudu as well.
>
> From my perspective, it would be great if you can share some details of
> your workload, especially if there are some areas you're finding Kudu
> lacking. Maybe we can spot some easy code changes we could make to improve
> performance, or suggest a tuning variable you could change.
>
> -Todd
>
>
> On May 27, 2016, at 9:19 PM, Todd Lipcon <t...@cloudera.com> wrote:
>
> On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim <bbuil...@gmail.com> wrote:
> Hi Mike,
>
> First of all, thanks for the link. It looks like an interesting read. I
> checked that Aerospike is currently at version 3.8.2.3, and in the article,
> they are evaluating version 3.5.4. The main thing that impressed me was
> their claim that they can beat Cassandra and HBase by 8x for writing and
> 25x for reading. Their big claim to fame is that Aerospike can write 1M
> records per second with only 50 nodes. I wanted to see if this is real.
>
> 1M records per second on 50 nodes is pretty doable by Kudu as well,
> depending on the size of your records and the insertion order. I've been
> playing with a ~70 node cluster recently and seen 1M+ writes/second
> sustained, and bursting above 4M. These are 1KB rows with 11 columns, and
> with pretty old HDD-only nodes. I think newer flash-based nodes could do
> better.
>
>
> To answer your q

Re: Performance Question

[Benjamin Kim]

Over the weekend, the row count is up to <500M. I will give it another few days 
to get to 1B rows. I still get consistent times ~15s for doing row counts 
despite the amount of data growing.

On another note, I got a solicitation email from SnappyData to evaluate their 
product. They claim to be the “Spark Data Store” with tight integration with 
Spark executors. It claims to be an OLTP and OLAP system with being an 
in-memory data store first then to disk. After going to several Spark events, 
it would seem that this is the new “hot” area for vendors. They all (MemSQL, 
Redis, Aerospike, Datastax, etc.) claim to be the best "Spark Data Store”. I’m 
wondering if Kudu will become this too? With the performance I’ve seen so far, 
it would seem that it can be a contender. All that is needed is a hardened 
Spark connector package, I would think. The next evaluation I will be 
conducting is to see if SnappyData’s claims are valid by doing my own tests.

Cheers,
Ben


> On Jun 15, 2016, at 12:47 AM, Todd Lipcon  wrote:
> 
> Hi Benjamin,
> 
> What workload are you using for benchmarks? Using spark or something more 
> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and 
> some queries
> 
> Todd
> 
> Todd
> 
> On Jun 15, 2016 8:10 AM, "Benjamin Kim"  > wrote:
> Hi Todd,
> 
> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am impressed. 
> Compared to HBase, read and write performance are better. Write performance 
> has the greatest improvement (> 4x), while read is > 1.5x. Albeit, these are 
> only preliminary tests. Do you know of a way to really do some conclusive 
> tests? I want to see if I can match your results on my 50 node cluster.
> 
> Thanks,
> Ben
> 
>> On May 30, 2016, at 10:33 AM, Todd Lipcon > > wrote:
>> 
>> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim > > wrote:
>> Todd,
>> 
>> It sounds like Kudu can possibly top or match those numbers put out by 
>> Aerospike. Do you have any performance statistics published or any 
>> instructions as to measure them myself as good way to test? In addition, 
>> this will be a test using Spark, so should I wait for Kudu version 0.9.0 
>> where support will be built in?
>> 
>> We don't have a lot of benchmarks published yet, especially on the write 
>> side. I've found that thorough cross-system benchmarks are very difficult to 
>> do fairly and accurately, and often times users end up misguided if they pay 
>> too much attention to them :) So, given a finite number of developers 
>> working on Kudu, I think we've tended to spend more time on the project 
>> itself and less time focusing on "competition". I'm sure there are use cases 
>> where Kudu will beat out Aerospike, and probably use cases where Aerospike 
>> will beat Kudu as well.
>> 
>> From my perspective, it would be great if you can share some details of your 
>> workload, especially if there are some areas you're finding Kudu lacking. 
>> Maybe we can spot some easy code changes we could make to improve 
>> performance, or suggest a tuning variable you could change.
>> 
>> -Todd
>> 
>> 
>>> On May 27, 2016, at 9:19 PM, Todd Lipcon >> > wrote:
>>> 
>>> On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim >> > wrote:
>>> Hi Mike,
>>> 
>>> First of all, thanks for the link. It looks like an interesting read. I 
>>> checked that Aerospike is currently at version 3.8.2.3, and in the article, 
>>> they are evaluating version 3.5.4. The main thing that impressed me was 
>>> their claim that they can beat Cassandra and HBase by 8x for writing and 
>>> 25x for reading. Their big claim to fame is that Aerospike can write 1M 
>>> records per second with only 50 nodes. I wanted to see if this is real.
>>> 
>>> 1M records per second on 50 nodes is pretty doable by Kudu as well, 
>>> depending on the size of your records and the insertion order. I've been 
>>> playing with a ~70 node cluster recently and seen 1M+ writes/second 
>>> sustained, and bursting above 4M. These are 1KB rows with 11 columns, and 
>>> with pretty old HDD-only nodes. I think newer flash-based nodes could do 
>>> better.
>>>  
>>> 
>>> To answer your questions, we have a DMP with user profiles with many 
>>> attributes. We create segmentation information off of these attributes to 
>>> classify them. Then, we can target advertising appropriately for our sales 
>>> department. Much of the data processing is for applying models on all or if 
>>> not most of every profile’s attributes to find similarities (nearest 
>>> neighbor/clustering) over a large number of rows when batch processing or a 
>>> small subset of rows for quick online scoring. So, our use case is a 
>>> typical advanced analytics scenario. We have tried HBase, but it doesn’t 
>>> work well 

Re: Performance Question

[Todd Lipcon]

On Thu, Jun 30, 2016 at 5:39 PM, Benjamin Kim  wrote:

> Hi Todd,
>
> I changed the key to be what you suggested, and I can’t tell the
> difference since it was already fast. But, I did get more numbers.
>

Yea, you won't see a substantial difference until you're inserting billions
of rows, etc, and the keys and/or bloom filters no longer fit in cache.


>
> > 104M rows in Kudu table
> - read: 8s
> - count: 16s
> - aggregate: 9s
>
> The time to read took much longer from 0.2s to 8s, counts were the same
> 16s, and aggregate queries look longer from 6s to 9s.
>

> I’m still impressed.
>

We aim to please ;-) If you have any interest in writing up these
experiments as a blog post, would be cool to post them for others to learn
from.

-Todd


> On Jun 15, 2016, at 12:47 AM, Todd Lipcon  wrote:
>
> Hi Benjamin,
>
> What workload are you using for benchmarks? Using spark or something more
> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and
> some queries
>
> Todd
>
> Todd
> On Jun 15, 2016 8:10 AM, "Benjamin Kim"  wrote:
>
>> Hi Todd,
>>
>> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am
>> impressed. Compared to HBase, read and write performance are better. Write
>> performance has the greatest improvement (> 4x), while read is > 1.5x.
>> Albeit, these are only preliminary tests. Do you know of a way to really do
>> some conclusive tests? I want to see if I can match your results on my 50
>> node cluster.
>>
>> Thanks,
>> Ben
>>
>> On May 30, 2016, at 10:33 AM, Todd Lipcon  wrote:
>>
>> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim  wrote:
>>
>>> Todd,
>>>
>>> It sounds like Kudu can possibly top or match those numbers put out by
>>> Aerospike. Do you have any performance statistics published or any
>>> instructions as to measure them myself as good way to test? In addition,
>>> this will be a test using Spark, so should I wait for Kudu version 0.9.0
>>> where support will be built in?
>>>
>>
>> We don't have a lot of benchmarks published yet, especially on the write
>> side. I've found that thorough cross-system benchmarks are very difficult
>> to do fairly and accurately, and often times users end up misguided if they
>> pay too much attention to them :) So, given a finite number of developers
>> working on Kudu, I think we've tended to spend more time on the project
>> itself and less time focusing on "competition". I'm sure there are use
>> cases where Kudu will beat out Aerospike, and probably use cases where
>> Aerospike will beat Kudu as well.
>>
>> From my perspective, it would be great if you can share some details of
>> your workload, especially if there are some areas you're finding Kudu
>> lacking. Maybe we can spot some easy code changes we could make to improve
>> performance, or suggest a tuning variable you could change.
>>
>> -Todd
>>
>>
>>> On May 27, 2016, at 9:19 PM, Todd Lipcon  wrote:
>>>
>>> On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim 
>>> wrote:
>>>
 Hi Mike,

 First of all, thanks for the link. It looks like an interesting read. I
 checked that Aerospike is currently at version 3.8.2.3, and in the article,
 they are evaluating version 3.5.4. The main thing that impressed me was
 their claim that they can beat Cassandra and HBase by 8x for writing and
 25x for reading. Their big claim to fame is that Aerospike can write 1M
 records per second with only 50 nodes. I wanted to see if this is real.

>>>
>>> 1M records per second on 50 nodes is pretty doable by Kudu as well,
>>> depending on the size of your records and the insertion order. I've been
>>> playing with a ~70 node cluster recently and seen 1M+ writes/second
>>> sustained, and bursting above 4M. These are 1KB rows with 11 columns, and
>>> with pretty old HDD-only nodes. I think newer flash-based nodes could do
>>> better.
>>>
>>>

 To answer your questions, we have a DMP with user profiles with many
 attributes. We create segmentation information off of these attributes to
 classify them. Then, we can target advertising appropriately for our sales
 department. Much of the data processing is for applying models on all or if
 not most of every profile’s attributes to find similarities (nearest
 neighbor/clustering) over a large number of rows when batch processing or a
 small subset of rows for quick online scoring. So, our use case is a
 typical advanced analytics scenario. We have tried HBase, but it doesn’t
 work well for these types of analytics.

 I read, that Aerospike in the release notes, they did do many
 improvements for batch and scan operations.

 I wonder what your thoughts are for using Kudu for this.

>>>
>>> Sounds like a good Kudu use case to me. I've heard great things about
>>> Aerospike for the low latency random access portion, but 

Re: Performance Question

[Benjamin Kim]

Hi Todd,

I changed the key to be what you suggested, and I can’t tell the difference 
since it was already fast. But, I did get more numbers.

> 104M rows in Kudu table
- read: 8s
- count: 16s
- aggregate: 9s

The time to read took much longer from 0.2s to 8s, counts were the same 16s, 
and aggregate queries look longer from 6s to 9s.

I’m still impressed.

Cheers,
Ben 

> On Jun 15, 2016, at 12:47 AM, Todd Lipcon  wrote:
> 
> Hi Benjamin,
> 
> What workload are you using for benchmarks? Using spark or something more 
> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and 
> some queries
> 
> Todd
> 
> Todd
> 
> On Jun 15, 2016 8:10 AM, "Benjamin Kim"  > wrote:
> Hi Todd,
> 
> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am impressed. 
> Compared to HBase, read and write performance are better. Write performance 
> has the greatest improvement (> 4x), while read is > 1.5x. Albeit, these are 
> only preliminary tests. Do you know of a way to really do some conclusive 
> tests? I want to see if I can match your results on my 50 node cluster.
> 
> Thanks,
> Ben
> 
>> On May 30, 2016, at 10:33 AM, Todd Lipcon > > wrote:
>> 
>> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim > > wrote:
>> Todd,
>> 
>> It sounds like Kudu can possibly top or match those numbers put out by 
>> Aerospike. Do you have any performance statistics published or any 
>> instructions as to measure them myself as good way to test? In addition, 
>> this will be a test using Spark, so should I wait for Kudu version 0.9.0 
>> where support will be built in?
>> 
>> We don't have a lot of benchmarks published yet, especially on the write 
>> side. I've found that thorough cross-system benchmarks are very difficult to 
>> do fairly and accurately, and often times users end up misguided if they pay 
>> too much attention to them :) So, given a finite number of developers 
>> working on Kudu, I think we've tended to spend more time on the project 
>> itself and less time focusing on "competition". I'm sure there are use cases 
>> where Kudu will beat out Aerospike, and probably use cases where Aerospike 
>> will beat Kudu as well.
>> 
>> From my perspective, it would be great if you can share some details of your 
>> workload, especially if there are some areas you're finding Kudu lacking. 
>> Maybe we can spot some easy code changes we could make to improve 
>> performance, or suggest a tuning variable you could change.
>> 
>> -Todd
>> 
>> 
>>> On May 27, 2016, at 9:19 PM, Todd Lipcon >> > wrote:
>>> 
>>> On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim >> > wrote:
>>> Hi Mike,
>>> 
>>> First of all, thanks for the link. It looks like an interesting read. I 
>>> checked that Aerospike is currently at version 3.8.2.3, and in the article, 
>>> they are evaluating version 3.5.4. The main thing that impressed me was 
>>> their claim that they can beat Cassandra and HBase by 8x for writing and 
>>> 25x for reading. Their big claim to fame is that Aerospike can write 1M 
>>> records per second with only 50 nodes. I wanted to see if this is real.
>>> 
>>> 1M records per second on 50 nodes is pretty doable by Kudu as well, 
>>> depending on the size of your records and the insertion order. I've been 
>>> playing with a ~70 node cluster recently and seen 1M+ writes/second 
>>> sustained, and bursting above 4M. These are 1KB rows with 11 columns, and 
>>> with pretty old HDD-only nodes. I think newer flash-based nodes could do 
>>> better.
>>>  
>>> 
>>> To answer your questions, we have a DMP with user profiles with many 
>>> attributes. We create segmentation information off of these attributes to 
>>> classify them. Then, we can target advertising appropriately for our sales 
>>> department. Much of the data processing is for applying models on all or if 
>>> not most of every profile’s attributes to find similarities (nearest 
>>> neighbor/clustering) over a large number of rows when batch processing or a 
>>> small subset of rows for quick online scoring. So, our use case is a 
>>> typical advanced analytics scenario. We have tried HBase, but it doesn’t 
>>> work well for these types of analytics.
>>> 
>>> I read, that Aerospike in the release notes, they did do many improvements 
>>> for batch and scan operations.
>>> 
>>> I wonder what your thoughts are for using Kudu for this.
>>> 
>>> Sounds like a good Kudu use case to me. I've heard great things about 
>>> Aerospike for the low latency random access portion, but I've also heard 
>>> that it's _very_ expensive, and not particularly suited to the columnar 
>>> scan workload. Lastly, I think the Apache license of Kudu is much more 
>>> appealing than the AGPL3 used by Aerospike. But, that's not really a 

Re: Performance Question

[Todd Lipcon]

On Wed, Jun 29, 2016 at 2:18 PM, Benjamin Kim  wrote:

> Todd,
>
> FYI. The key  is unique for every row so rows are not going to already
> exist. Basically, everything is an INSERT.
>
> val generateUUID = udf(() => UUID.randomUUID().toString)
>
> As you can see, we are using UUID java library to create the key.
>

OK. You will have better insert performance if instead your key is
something that is increasing with time (eg System.currentTimeMillis() +
UUID).

-Todd


> On Jun 29, 2016, at 1:32 PM, Todd Lipcon  wrote:
>
> On Wed, Jun 29, 2016 at 11:32 AM, Benjamin Kim  wrote:
>
>> Todd,
>>
>> I started Spark streaming more events into Kudu. Performance is great
>> there too! With HBase, it’s fast too, but I noticed that it pauses here and
>> there, making it take seconds for > 40k rows at a time, while Kudu doesn’t.
>> The progress bar just blinks by. I will keep this running until it hits 1B
>> rows and rerun my performance tests. This, hopefully, will give better
>> numbers.
>>
>
> Cool! We have invested a lot of work in making Kudu have consistent
> performance, like you mentioned. It's generally been my experience that
> most mature ops people would prefer a system which consistently performs
> well rather than one which has higher peak performance but occasionally
> stalls.
>
> BTW, what is your row key design? One exception to the above is that, if
> you're doing random inserts, you may see performance "fall off a cliff"
> once the size of your key columns becomes larger than the aggregate memory
> size of your cluster, if you're running on hard disks. Our inserts require
> checks for duplicate keys, and that can cause random disk IOs if your keys
> don't fit comfortably in cache. This is one area that HBase is
> fundamentally going to be faster based on its design.
>
> -Todd
>
>
>> On Jun 28, 2016, at 4:26 PM, Todd Lipcon  wrote:
>>
>> Cool, thanks for the report, Ben. For what it's worth, I think there's
>> still some low hanging fruit in the Spark connector for Kudu (for example,
>> I believe locality on reads is currently broken). So, you can expect
>> performance to continue to improve in future versions. I'd also be
>> interested to see results on Kudu for a much larger dataset - my guess is a
>> lot of the 6 seconds you're seeing is constant overhead from Spark job
>> setup, etc, given that the performance doesn't seem to get slower as you
>> went from 700K rows to 13M rows.
>>
>> -Todd
>>
>> On Tue, Jun 28, 2016 at 3:03 PM, Benjamin Kim  wrote:
>>
>>> FYI.
>>>
>>> I did a quick-n-dirty performance test.
>>>
>>> First, the setup:
>>> QA cluster:
>>>
>>>- 15 data nodes
>>>   - 64GB memory each
>>>   - HBase is using 4GB of memory
>>>   - Kudu is using 1GB of memory
>>>- 1 HBase/Kudu master node
>>>   - 64GB memory
>>>   - HBase/Kudu master is using 1GB of memory each
>>>- 10Gb Ethernet
>>>
>>>
>>> Using Spark on both to load/read events data (84 columns per row), I was
>>> able to record performance for each. On the HBase side, I used the Phoenix
>>> 4.7 Spark plugin where DataFrames can be used directly. On the Kudu side, I
>>> used the Spark connector. I created an events table in Phoenix using the
>>> CREATE TABLE statement and created the equivalent in Kudu using the Spark
>>> method based off of a DataFrame schema.
>>>
>>> Here are the numbers for Phoenix/HBase.
>>> 1st run:
>>> > 715k rows
>>> - write: 2.7m
>>>
>>> > 715k rows in HBase table
>>> - read: 0.1s
>>> - count: 3.8s
>>> - aggregate: 61s
>>>
>>> 2nd run:
>>> > 5.2M rows
>>> - write: 11m
>>> * had 4 region servers go down, had to retry the 5.2M row write
>>>
>>> > 5.9M rows in HBase table
>>> - read: 8s
>>> - count: 3m
>>> - aggregate: 46s
>>>
>>> 3rd run:
>>> > 6.8M rows
>>> - write: 9.6m
>>>
>>> > 12.7M rows
>>> - read: 10s
>>> - count: 3m
>>> - aggregate: 44s
>>>
>>>
>>> Here are the numbers for Kudu.
>>> 1st run:
>>> > 715k rows
>>> - write: 18s
>>>
>>> > 715k rows in Kudu table
>>> - read: 0.2s
>>> - count: 18s
>>> - aggregate: 5s
>>>
>>> 2nd run:
>>> > 5.2M rows
>>> - write: 33s
>>>
>>> > 5.9M rows in Kudu table
>>> - read: 0.2s
>>> - count: 16s
>>> - aggregate: 6s
>>>
>>> 3rd run:
>>> > 6.8M rows
>>> - write: 27s
>>>
>>> > 12.7M rows in Kudu table
>>> - read: 0.2s
>>> - count: 16s
>>> - aggregate: 6s
>>>
>>> The Kudu results are impressive if you take these number as-is. Kudu is
>>> close to 18x faster at writing (UPSERT). Kudu is 30x faster at reading
>>> (HBase times increase as data size grows).  Kudu is 7x faster at full row
>>> counts. Lastly, Kudu is 3x faster doing an aggregate query (count distinct
>>> event_id’s per user_id). *Remember that this is small cluster, times are
>>> still respectable for both systems, HBase could have been configured
>>> better, and the HBase table could have been better tuned.
>>>
>>> Cheers,
>>> Ben
>>>
>>>
>>> On Jun 15, 2016, at 10:13 AM, Dan Burkert 

Re: Performance Question

[Benjamin Kim]

Todd,

FYI. The key  is unique for every row so rows are not going to already exist. 
Basically, everything is an INSERT.

val generateUUID = udf(() => UUID.randomUUID().toString)

As you can see, we are using UUID java library to create the key.

Cheers,
Ben

> On Jun 29, 2016, at 1:32 PM, Todd Lipcon  wrote:
> 
> On Wed, Jun 29, 2016 at 11:32 AM, Benjamin Kim  > wrote:
> Todd,
> 
> I started Spark streaming more events into Kudu. Performance is great there 
> too! With HBase, it’s fast too, but I noticed that it pauses here and there, 
> making it take seconds for > 40k rows at a time, while Kudu doesn’t. The 
> progress bar just blinks by. I will keep this running until it hits 1B rows 
> and rerun my performance tests. This, hopefully, will give better numbers.
> 
> Cool! We have invested a lot of work in making Kudu have consistent 
> performance, like you mentioned. It's generally been my experience that most 
> mature ops people would prefer a system which consistently performs well 
> rather than one which has higher peak performance but occasionally stalls.
> 
> BTW, what is your row key design? One exception to the above is that, if 
> you're doing random inserts, you may see performance "fall off a cliff" once 
> the size of your key columns becomes larger than the aggregate memory size of 
> your cluster, if you're running on hard disks. Our inserts require checks for 
> duplicate keys, and that can cause random disk IOs if your keys don't fit 
> comfortably in cache. This is one area that HBase is fundamentally going to 
> be faster based on its design.
> 
> -Todd
> 
> 
>> On Jun 28, 2016, at 4:26 PM, Todd Lipcon > > wrote:
>> 
>> Cool, thanks for the report, Ben. For what it's worth, I think there's still 
>> some low hanging fruit in the Spark connector for Kudu (for example, I 
>> believe locality on reads is currently broken). So, you can expect 
>> performance to continue to improve in future versions. I'd also be 
>> interested to see results on Kudu for a much larger dataset - my guess is a 
>> lot of the 6 seconds you're seeing is constant overhead from Spark job 
>> setup, etc, given that the performance doesn't seem to get slower as you 
>> went from 700K rows to 13M rows.
>> 
>> -Todd
>> 
>> On Tue, Jun 28, 2016 at 3:03 PM, Benjamin Kim > > wrote:
>> FYI.
>> 
>> I did a quick-n-dirty performance test.
>> 
>> First, the setup:
>> QA cluster:
>> 15 data nodes
>> 64GB memory each
>> HBase is using 4GB of memory
>> Kudu is using 1GB of memory
>> 1 HBase/Kudu master node
>> 64GB memory
>> HBase/Kudu master is using 1GB of memory each
>> 10Gb Ethernet
>> 
>> Using Spark on both to load/read events data (84 columns per row), I was 
>> able to record performance for each. On the HBase side, I used the Phoenix 
>> 4.7 Spark plugin where DataFrames can be used directly. On the Kudu side, I 
>> used the Spark connector. I created an events table in Phoenix using the 
>> CREATE TABLE statement and created the equivalent in Kudu using the Spark 
>> method based off of a DataFrame schema.
>> 
>> Here are the numbers for Phoenix/HBase.
>> 1st run:
>> > 715k rows
>> - write: 2.7m
>> 
>> > 715k rows in HBase table
>> - read: 0.1s
>> - count: 3.8s
>> - aggregate: 61s
>> 
>> 2nd run:
>> > 5.2M rows
>> - write: 11m
>> * had 4 region servers go down, had to retry the 5.2M row write
>> 
>> > 5.9M rows in HBase table
>> - read: 8s
>> - count: 3m
>> - aggregate: 46s
>> 
>> 3rd run:
>> > 6.8M rows
>> - write: 9.6m
>> 
>> > 12.7M rows
>> - read: 10s
>> - count: 3m
>> - aggregate: 44s
>> 
>> 
>> Here are the numbers for Kudu.
>> 1st run:
>> > 715k rows
>> - write: 18s
>> 
>> > 715k rows in Kudu table
>> - read: 0.2s
>> - count: 18s
>> - aggregate: 5s
>> 
>> 2nd run:
>> > 5.2M rows
>> - write: 33s
>> 
>> > 5.9M rows in Kudu table
>> - read: 0.2s
>> - count: 16s
>> - aggregate: 6s
>> 
>> 3rd run:
>> > 6.8M rows
>> - write: 27s
>> 
>> > 12.7M rows in Kudu table
>> - read: 0.2s
>> - count: 16s
>> - aggregate: 6s
>> 
>> The Kudu results are impressive if you take these number as-is. Kudu is 
>> close to 18x faster at writing (UPSERT). Kudu is 30x faster at reading 
>> (HBase times increase as data size grows).  Kudu is 7x faster at full row 
>> counts. Lastly, Kudu is 3x faster doing an aggregate query (count distinct 
>> event_id’s per user_id). *Remember that this is small cluster, times are 
>> still respectable for both systems, HBase could have been configured better, 
>> and the HBase table could have been better tuned.
>> 
>> Cheers,
>> Ben
>> 
>> 
>>> On Jun 15, 2016, at 10:13 AM, Dan Burkert >> > wrote:
>>> 
>>> Adding partition splits when range partitioning is done via the 
>>> CreateTableOptions.addSplitRow 
>>> 

Re: Performance Question

[Todd Lipcon]

On Wed, Jun 29, 2016 at 11:32 AM, Benjamin Kim  wrote:

> Todd,
>
> I started Spark streaming more events into Kudu. Performance is great
> there too! With HBase, it’s fast too, but I noticed that it pauses here and
> there, making it take seconds for > 40k rows at a time, while Kudu doesn’t.
> The progress bar just blinks by. I will keep this running until it hits 1B
> rows and rerun my performance tests. This, hopefully, will give better
> numbers.
>

Cool! We have invested a lot of work in making Kudu have consistent
performance, like you mentioned. It's generally been my experience that
most mature ops people would prefer a system which consistently performs
well rather than one which has higher peak performance but occasionally
stalls.

BTW, what is your row key design? One exception to the above is that, if
you're doing random inserts, you may see performance "fall off a cliff"
once the size of your key columns becomes larger than the aggregate memory
size of your cluster, if you're running on hard disks. Our inserts require
checks for duplicate keys, and that can cause random disk IOs if your keys
don't fit comfortably in cache. This is one area that HBase is
fundamentally going to be faster based on its design.

-Todd


> On Jun 28, 2016, at 4:26 PM, Todd Lipcon  wrote:
>
> Cool, thanks for the report, Ben. For what it's worth, I think there's
> still some low hanging fruit in the Spark connector for Kudu (for example,
> I believe locality on reads is currently broken). So, you can expect
> performance to continue to improve in future versions. I'd also be
> interested to see results on Kudu for a much larger dataset - my guess is a
> lot of the 6 seconds you're seeing is constant overhead from Spark job
> setup, etc, given that the performance doesn't seem to get slower as you
> went from 700K rows to 13M rows.
>
> -Todd
>
> On Tue, Jun 28, 2016 at 3:03 PM, Benjamin Kim  wrote:
>
>> FYI.
>>
>> I did a quick-n-dirty performance test.
>>
>> First, the setup:
>> QA cluster:
>>
>>- 15 data nodes
>>   - 64GB memory each
>>   - HBase is using 4GB of memory
>>   - Kudu is using 1GB of memory
>>- 1 HBase/Kudu master node
>>   - 64GB memory
>>   - HBase/Kudu master is using 1GB of memory each
>>- 10Gb Ethernet
>>
>>
>> Using Spark on both to load/read events data (84 columns per row), I was
>> able to record performance for each. On the HBase side, I used the Phoenix
>> 4.7 Spark plugin where DataFrames can be used directly. On the Kudu side, I
>> used the Spark connector. I created an events table in Phoenix using the
>> CREATE TABLE statement and created the equivalent in Kudu using the Spark
>> method based off of a DataFrame schema.
>>
>> Here are the numbers for Phoenix/HBase.
>> 1st run:
>> > 715k rows
>> - write: 2.7m
>>
>> > 715k rows in HBase table
>> - read: 0.1s
>> - count: 3.8s
>> - aggregate: 61s
>>
>> 2nd run:
>> > 5.2M rows
>> - write: 11m
>> * had 4 region servers go down, had to retry the 5.2M row write
>>
>> > 5.9M rows in HBase table
>> - read: 8s
>> - count: 3m
>> - aggregate: 46s
>>
>> 3rd run:
>> > 6.8M rows
>> - write: 9.6m
>>
>> > 12.7M rows
>> - read: 10s
>> - count: 3m
>> - aggregate: 44s
>>
>>
>> Here are the numbers for Kudu.
>> 1st run:
>> > 715k rows
>> - write: 18s
>>
>> > 715k rows in Kudu table
>> - read: 0.2s
>> - count: 18s
>> - aggregate: 5s
>>
>> 2nd run:
>> > 5.2M rows
>> - write: 33s
>>
>> > 5.9M rows in Kudu table
>> - read: 0.2s
>> - count: 16s
>> - aggregate: 6s
>>
>> 3rd run:
>> > 6.8M rows
>> - write: 27s
>>
>> > 12.7M rows in Kudu table
>> - read: 0.2s
>> - count: 16s
>> - aggregate: 6s
>>
>> The Kudu results are impressive if you take these number as-is. Kudu is
>> close to 18x faster at writing (UPSERT). Kudu is 30x faster at reading
>> (HBase times increase as data size grows).  Kudu is 7x faster at full row
>> counts. Lastly, Kudu is 3x faster doing an aggregate query (count distinct
>> event_id’s per user_id). *Remember that this is small cluster, times are
>> still respectable for both systems, HBase could have been configured
>> better, and the HBase table could have been better tuned.
>>
>> Cheers,
>> Ben
>>
>>
>> On Jun 15, 2016, at 10:13 AM, Dan Burkert  wrote:
>>
>> Adding partition splits when range partitioning is done via the
>> CreateTableOptions.addSplitRow
>> 
>>  method.
>> You can find more about the different partitioning options in the schema
>> design guide
>> .  We
>> generally recommend sticking to hash partitioning if possible, since you
>> don't have to determine your own split rows.
>>
>> - Dan
>>
>> On Wed, Jun 15, 2016 at 9:17 AM, Benjamin Kim  wrote:
>>
>>> Todd,
>>>
>>> I think the locality is not within our setup. We have the compute

Re: Performance Question

[Benjamin Kim]

Todd,

I started Spark streaming more events into Kudu. Performance is great there 
too! With HBase, it’s fast too, but I noticed that it pauses here and there, 
making it take seconds for > 40k rows at a time, while Kudu doesn’t. The 
progress bar just blinks by. I will keep this running until it hits 1B rows and 
rerun my performance tests. This, hopefully, will give better numbers.

Thanks,
Ben


> On Jun 28, 2016, at 4:26 PM, Todd Lipcon  wrote:
> 
> Cool, thanks for the report, Ben. For what it's worth, I think there's still 
> some low hanging fruit in the Spark connector for Kudu (for example, I 
> believe locality on reads is currently broken). So, you can expect 
> performance to continue to improve in future versions. I'd also be interested 
> to see results on Kudu for a much larger dataset - my guess is a lot of the 6 
> seconds you're seeing is constant overhead from Spark job setup, etc, given 
> that the performance doesn't seem to get slower as you went from 700K rows to 
> 13M rows.
> 
> -Todd
> 
> On Tue, Jun 28, 2016 at 3:03 PM, Benjamin Kim  > wrote:
> FYI.
> 
> I did a quick-n-dirty performance test.
> 
> First, the setup:
> QA cluster:
> 15 data nodes
> 64GB memory each
> HBase is using 4GB of memory
> Kudu is using 1GB of memory
> 1 HBase/Kudu master node
> 64GB memory
> HBase/Kudu master is using 1GB of memory each
> 10Gb Ethernet
> 
> Using Spark on both to load/read events data (84 columns per row), I was able 
> to record performance for each. On the HBase side, I used the Phoenix 4.7 
> Spark plugin where DataFrames can be used directly. On the Kudu side, I used 
> the Spark connector. I created an events table in Phoenix using the CREATE 
> TABLE statement and created the equivalent in Kudu using the Spark method 
> based off of a DataFrame schema.
> 
> Here are the numbers for Phoenix/HBase.
> 1st run:
> > 715k rows
> - write: 2.7m
> 
> > 715k rows in HBase table
> - read: 0.1s
> - count: 3.8s
> - aggregate: 61s
> 
> 2nd run:
> > 5.2M rows
> - write: 11m
> * had 4 region servers go down, had to retry the 5.2M row write
> 
> > 5.9M rows in HBase table
> - read: 8s
> - count: 3m
> - aggregate: 46s
> 
> 3rd run:
> > 6.8M rows
> - write: 9.6m
> 
> > 12.7M rows
> - read: 10s
> - count: 3m
> - aggregate: 44s
> 
> 
> Here are the numbers for Kudu.
> 1st run:
> > 715k rows
> - write: 18s
> 
> > 715k rows in Kudu table
> - read: 0.2s
> - count: 18s
> - aggregate: 5s
> 
> 2nd run:
> > 5.2M rows
> - write: 33s
> 
> > 5.9M rows in Kudu table
> - read: 0.2s
> - count: 16s
> - aggregate: 6s
> 
> 3rd run:
> > 6.8M rows
> - write: 27s
> 
> > 12.7M rows in Kudu table
> - read: 0.2s
> - count: 16s
> - aggregate: 6s
> 
> The Kudu results are impressive if you take these number as-is. Kudu is close 
> to 18x faster at writing (UPSERT). Kudu is 30x faster at reading (HBase times 
> increase as data size grows).  Kudu is 7x faster at full row counts. Lastly, 
> Kudu is 3x faster doing an aggregate query (count distinct event_id’s per 
> user_id). *Remember that this is small cluster, times are still respectable 
> for both systems, HBase could have been configured better, and the HBase 
> table could have been better tuned.
> 
> Cheers,
> Ben
> 
> 
>> On Jun 15, 2016, at 10:13 AM, Dan Burkert > > wrote:
>> 
>> Adding partition splits when range partitioning is done via the 
>> CreateTableOptions.addSplitRow 
>> 
>>  method.  You can find more about the different partitioning options in the 
>> schema design guide 
>> .  We generally 
>> recommend sticking to hash partitioning if possible, since you don't have to 
>> determine your own split rows.
>> 
>> - Dan
>> 
>> On Wed, Jun 15, 2016 at 9:17 AM, Benjamin Kim > > wrote:
>> Todd,
>> 
>> I think the locality is not within our setup. We have the compute cluster 
>> with Spark, YARN, etc. on its own, and we have the storage cluster with 
>> HBase, Kudu, etc. on another. We beefed up the hardware specs on the compute 
>> cluster and beefed up storage capacity on the storage cluster. We got this 
>> setup idea from the Databricks folks. I do have a question. I created the 
>> table to use range partition on columns. I see that if I use hash partition 
>> I can set the number of splits, but how do I do that using range (50 nodes * 
>> 10 = 500 splits)?
>> 
>> Thanks,
>> Ben
>> 
>> 
>>> On Jun 15, 2016, at 9:11 AM, Todd Lipcon >> > wrote:
>>> 
>>> Awesome use case. One thing to keep in mind is that spark parallelism will 
>>> be limited by the number of tablets. So, you might want to split into 10 or 
>>> so buckets per node to get the best query throughput.
>>> 
>>> Usually if 

Re: Performance Question

[Todd Lipcon]

Cool, thanks for the report, Ben. For what it's worth, I think there's
still some low hanging fruit in the Spark connector for Kudu (for example,
I believe locality on reads is currently broken). So, you can expect
performance to continue to improve in future versions. I'd also be
interested to see results on Kudu for a much larger dataset - my guess is a
lot of the 6 seconds you're seeing is constant overhead from Spark job
setup, etc, given that the performance doesn't seem to get slower as you
went from 700K rows to 13M rows.

-Todd

On Tue, Jun 28, 2016 at 3:03 PM, Benjamin Kim  wrote:

> FYI.
>
> I did a quick-n-dirty performance test.
>
> First, the setup:
> QA cluster:
>
>- 15 data nodes
>   - 64GB memory each
>   - HBase is using 4GB of memory
>   - Kudu is using 1GB of memory
>- 1 HBase/Kudu master node
>   - 64GB memory
>   - HBase/Kudu master is using 1GB of memory each
>- 10Gb Ethernet
>
>
> Using Spark on both to load/read events data (84 columns per row), I was
> able to record performance for each. On the HBase side, I used the Phoenix
> 4.7 Spark plugin where DataFrames can be used directly. On the Kudu side, I
> used the Spark connector. I created an events table in Phoenix using the
> CREATE TABLE statement and created the equivalent in Kudu using the Spark
> method based off of a DataFrame schema.
>
> Here are the numbers for Phoenix/HBase.
> 1st run:
> > 715k rows
> - write: 2.7m
>
> > 715k rows in HBase table
> - read: 0.1s
> - count: 3.8s
> - aggregate: 61s
>
> 2nd run:
> > 5.2M rows
> - write: 11m
> * had 4 region servers go down, had to retry the 5.2M row write
>
> > 5.9M rows in HBase table
> - read: 8s
> - count: 3m
> - aggregate: 46s
>
> 3rd run:
> > 6.8M rows
> - write: 9.6m
>
> > 12.7M rows
> - read: 10s
> - count: 3m
> - aggregate: 44s
>
>
> Here are the numbers for Kudu.
> 1st run:
> > 715k rows
> - write: 18s
>
> > 715k rows in Kudu table
> - read: 0.2s
> - count: 18s
> - aggregate: 5s
>
> 2nd run:
> > 5.2M rows
> - write: 33s
>
> > 5.9M rows in Kudu table
> - read: 0.2s
> - count: 16s
> - aggregate: 6s
>
> 3rd run:
> > 6.8M rows
> - write: 27s
>
> > 12.7M rows in Kudu table
> - read: 0.2s
> - count: 16s
> - aggregate: 6s
>
> The Kudu results are impressive if you take these number as-is. Kudu is
> close to 18x faster at writing (UPSERT). Kudu is 30x faster at reading
> (HBase times increase as data size grows).  Kudu is 7x faster at full row
> counts. Lastly, Kudu is 3x faster doing an aggregate query (count distinct
> event_id’s per user_id). *Remember that this is small cluster, times are
> still respectable for both systems, HBase could have been configured
> better, and the HBase table could have been better tuned.
>
> Cheers,
> Ben
>
>
> On Jun 15, 2016, at 10:13 AM, Dan Burkert  wrote:
>
> Adding partition splits when range partitioning is done via the
> CreateTableOptions.addSplitRow
> 
>  method.
> You can find more about the different partitioning options in the schema
> design guide .
> We generally recommend sticking to hash partitioning if possible, since you
> don't have to determine your own split rows.
>
> - Dan
>
> On Wed, Jun 15, 2016 at 9:17 AM, Benjamin Kim  wrote:
>
>> Todd,
>>
>> I think the locality is not within our setup. We have the compute cluster
>> with Spark, YARN, etc. on its own, and we have the storage cluster with
>> HBase, Kudu, etc. on another. We beefed up the hardware specs on the
>> compute cluster and beefed up storage capacity on the storage cluster. We
>> got this setup idea from the Databricks folks. I do have a question. I
>> created the table to use range partition on columns. I see that if I use
>> hash partition I can set the number of splits, but how do I do that using
>> range (50 nodes * 10 = 500 splits)?
>>
>> Thanks,
>> Ben
>>
>>
>> On Jun 15, 2016, at 9:11 AM, Todd Lipcon  wrote:
>>
>> Awesome use case. One thing to keep in mind is that spark parallelism
>> will be limited by the number of tablets. So, you might want to split into
>> 10 or so buckets per node to get the best query throughput.
>>
>> Usually if you run top on some machines while running the query you can
>> see if it is fully utilizing the cores.
>>
>> Another known issue right now is that spark locality isn't working
>> properly on replicated tables so you will use a lot of network traffic. For
>> a perf test you might want to try a table with replication count 1
>> On Jun 15, 2016 5:26 PM, "Benjamin Kim"  wrote:
>>
>> Hi Todd,
>>
>> I did a simple test of our ad events. We stream using Spark Streaming
>> directly into HBase, and the Data Analysts/Scientists do some
>> insight/discovery work plus some reports generation. For the reports, we
>> use SQL, and the 

Re: Performance Question

[Benjamin Kim]

FYI.

I did a quick-n-dirty performance test.

First, the setup:
QA cluster:
15 data nodes
64GB memory each
HBase is using 4GB of memory
Kudu is using 1GB of memory
1 HBase/Kudu master node
64GB memory
HBase/Kudu master is using 1GB of memory each
10Gb Ethernet

Using Spark on both to load/read events data (84 columns per row), I was able 
to record performance for each. On the HBase side, I used the Phoenix 4.7 Spark 
plugin where DataFrames can be used directly. On the Kudu side, I used the 
Spark connector. I created an events table in Phoenix using the CREATE TABLE 
statement and created the equivalent in Kudu using the Spark method based off 
of a DataFrame schema.

Here are the numbers for Phoenix/HBase.
1st run:
> 715k rows
- write: 2.7m

> 715k rows in HBase table
- read: 0.1s
- count: 3.8s
- aggregate: 61s

2nd run:
> 5.2M rows
- write: 11m
* had 4 region servers go down, had to retry the 5.2M row write

> 5.9M rows in HBase table
- read: 8s
- count: 3m
- aggregate: 46s

3rd run:
> 6.8M rows
- write: 9.6m

> 12.7M rows
- read: 10s
- count: 3m
- aggregate: 44s


Here are the numbers for Kudu.
1st run:
> 715k rows
- write: 18s

> 715k rows in Kudu table
- read: 0.2s
- count: 18s
- aggregate: 5s

2nd run:
> 5.2M rows
- write: 33s

> 5.9M rows in Kudu table
- read: 0.2s
- count: 16s
- aggregate: 6s

3rd run:
> 6.8M rows
- write: 27s

> 12.7M rows in Kudu table
- read: 0.2s
- count: 16s
- aggregate: 6s

The Kudu results are impressive if you take these number as-is. Kudu is close 
to 18x faster at writing (UPSERT). Kudu is 30x faster at reading (HBase times 
increase as data size grows).  Kudu is 7x faster at full row counts. Lastly, 
Kudu is 3x faster doing an aggregate query (count distinct event_id’s per 
user_id). *Remember that this is small cluster, times are still respectable for 
both systems, HBase could have been configured better, and the HBase table 
could have been better tuned.

Cheers,
Ben


> On Jun 15, 2016, at 10:13 AM, Dan Burkert  wrote:
> 
> Adding partition splits when range partitioning is done via the 
> CreateTableOptions.addSplitRow 
> 
>  method.  You can find more about the different partitioning options in the 
> schema design guide 
> .  We generally 
> recommend sticking to hash partitioning if possible, since you don't have to 
> determine your own split rows.
> 
> - Dan
> 
> On Wed, Jun 15, 2016 at 9:17 AM, Benjamin Kim  > wrote:
> Todd,
> 
> I think the locality is not within our setup. We have the compute cluster 
> with Spark, YARN, etc. on its own, and we have the storage cluster with 
> HBase, Kudu, etc. on another. We beefed up the hardware specs on the compute 
> cluster and beefed up storage capacity on the storage cluster. We got this 
> setup idea from the Databricks folks. I do have a question. I created the 
> table to use range partition on columns. I see that if I use hash partition I 
> can set the number of splits, but how do I do that using range (50 nodes * 10 
> = 500 splits)?
> 
> Thanks,
> Ben
> 
> 
>> On Jun 15, 2016, at 9:11 AM, Todd Lipcon > > wrote:
>> 
>> Awesome use case. One thing to keep in mind is that spark parallelism will 
>> be limited by the number of tablets. So, you might want to split into 10 or 
>> so buckets per node to get the best query throughput.
>> 
>> Usually if you run top on some machines while running the query you can see 
>> if it is fully utilizing the cores.
>> 
>> Another known issue right now is that spark locality isn't working properly 
>> on replicated tables so you will use a lot of network traffic. For a perf 
>> test you might want to try a table with replication count 1
>> 
>> On Jun 15, 2016 5:26 PM, "Benjamin Kim" > > wrote:
>> Hi Todd,
>> 
>> I did a simple test of our ad events. We stream using Spark Streaming 
>> directly into HBase, and the Data Analysts/Scientists do some 
>> insight/discovery work plus some reports generation. For the reports, we use 
>> SQL, and the more deeper stuff, we use Spark. In Spark, our main data 
>> currency store of choice is DataFrames.
>> 
>> The schema is around 83 columns wide where most are of the string data type.
>> 
>> "event_type", "timestamp", "event_valid", "event_subtype", "user_ip", 
>> "user_id", "mappable_id",
>> "cookie_status", "profile_status", "user_status", "previous_timestamp", 
>> "user_agent", "referer",
>> "host_domain", "uri", "request_elapsed", "browser_languages", "acamp_id", 
>> "creative_id",
>> "location_id", “pcamp_id",
>> "pdomain_id", "continent_code", "country", "region", "dma", "city", "zip", 
>> "isp", "line_speed",
>> "gender", "year_of_birth", "behaviors_read", "behaviors_written", 
>> "key_value_pairs", 

Re: Performance Question

[Dan Burkert]

Adding partition splits when range partitioning is done via the
CreateTableOptions.addSplitRow

method.
You can find more about the different partitioning options in the schema
design guide .
We generally recommend sticking to hash partitioning if possible, since you
don't have to determine your own split rows.

- Dan

On Wed, Jun 15, 2016 at 9:17 AM, Benjamin Kim  wrote:

> Todd,
>
> I think the locality is not within our setup. We have the compute cluster
> with Spark, YARN, etc. on its own, and we have the storage cluster with
> HBase, Kudu, etc. on another. We beefed up the hardware specs on the
> compute cluster and beefed up storage capacity on the storage cluster. We
> got this setup idea from the Databricks folks. I do have a question. I
> created the table to use range partition on columns. I see that if I use
> hash partition I can set the number of splits, but how do I do that using
> range (50 nodes * 10 = 500 splits)?
>
> Thanks,
> Ben
>
>
> On Jun 15, 2016, at 9:11 AM, Todd Lipcon  wrote:
>
> Awesome use case. One thing to keep in mind is that spark parallelism will
> be limited by the number of tablets. So, you might want to split into 10 or
> so buckets per node to get the best query throughput.
>
> Usually if you run top on some machines while running the query you can
> see if it is fully utilizing the cores.
>
> Another known issue right now is that spark locality isn't working
> properly on replicated tables so you will use a lot of network traffic. For
> a perf test you might want to try a table with replication count 1
> On Jun 15, 2016 5:26 PM, "Benjamin Kim"  wrote:
>
> Hi Todd,
>
> I did a simple test of our ad events. We stream using Spark Streaming
> directly into HBase, and the Data Analysts/Scientists do some
> insight/discovery work plus some reports generation. For the reports, we
> use SQL, and the more deeper stuff, we use Spark. In Spark, our main data
> currency store of choice is DataFrames.
>
> The schema is around 83 columns wide where most are of the string data
> type.
>
> "event_type", "timestamp", "event_valid", "event_subtype", "user_ip",
> "user_id", "mappable_id",
> "cookie_status", "profile_status", "user_status", "previous_timestamp",
> "user_agent", "referer",
> "host_domain", "uri", "request_elapsed", "browser_languages", "acamp_id",
> "creative_id",
> "location_id", “pcamp_id",
> "pdomain_id", "continent_code", "country", "region", "dma", "city", "zip",
> "isp", "line_speed",
> "gender", "year_of_birth", "behaviors_read", "behaviors_written",
> "key_value_pairs", "acamp_candidates",
> "tag_format", "optimizer_name", "optimizer_version", "optimizer_ip",
> "pixel_id", “video_id",
> "video_network_id", "video_time_watched", "video_percentage_watched",
> "video_media_type",
> "video_player_iframed", "video_player_in_view", "video_player_width",
> "video_player_height",
> "conversion_valid_sale", "conversion_sale_amount",
> "conversion_commission_amount", "conversion_step",
> "conversion_currency", "conversion_attribution", "conversion_offer_id",
> "custom_info", "frequency",
> "recency_seconds", "cost", "revenue", “optimizer_acamp_id",
> "optimizer_creative_id", "optimizer_ecpm", "impression_id",
> "diagnostic_data",
> "user_profile_mapping_source", "latitude", "longitude", "area_code",
> "gmt_offset", "in_dst",
> "proxy_type", "mobile_carrier", "pop", "hostname", "profile_expires",
> "timestamp_iso", "reference_id",
> "identity_organization", "identity_method"
>
> Most queries are like counts of how many users use what browser, how many
> are unique users, etc. The part that scares most users is when it comes to
> joining this data with other dimension/3rd party events tables because of
> shear size of it.
>
> We do what most companies do, similar to what I saw in earlier
> presentations of Kudu. We dump data out of HBase into partitioned Parquet
> tables to make query performance manageable.
>
> I will coordinate with a data scientist today to do some tests. He is
> working on identity matching/record linking of users from 2 domains: US and
> Singapore, using probabilistic deduping algorithms. I will load the data
> from ad events from both countries, and let him run his process against
> this data in Kudu. I hope this will “wow” the team.
>
> Thanks,
> Ben
>
> On Jun 15, 2016, at 12:47 AM, Todd Lipcon  wrote:
>
> Hi Benjamin,
>
> What workload are you using for benchmarks? Using spark or something more
> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and
> some queries
>
> Todd
>
> Todd
> On Jun 15, 2016 8:10 AM, "Benjamin Kim"  wrote:
>
>> Hi Todd,
>>
>> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am
>> impressed. Compared to HBase, read and write 

Re: Performance Question

[Benjamin Kim]

Todd,

I think the locality is not within our setup. We have the compute cluster with 
Spark, YARN, etc. on its own, and we have the storage cluster with HBase, Kudu, 
etc. on another. We beefed up the hardware specs on the compute cluster and 
beefed up storage capacity on the storage cluster. We got this setup idea from 
the Databricks folks. I do have a question. I created the table to use range 
partition on columns. I see that if I use hash partition I can set the number 
of splits, but how do I do that using range (50 nodes * 10 = 500 splits)?

Thanks,
Ben

> On Jun 15, 2016, at 9:11 AM, Todd Lipcon  wrote:
> 
> Awesome use case. One thing to keep in mind is that spark parallelism will be 
> limited by the number of tablets. So, you might want to split into 10 or so 
> buckets per node to get the best query throughput.
> 
> Usually if you run top on some machines while running the query you can see 
> if it is fully utilizing the cores.
> 
> Another known issue right now is that spark locality isn't working properly 
> on replicated tables so you will use a lot of network traffic. For a perf 
> test you might want to try a table with replication count 1
> 
> On Jun 15, 2016 5:26 PM, "Benjamin Kim"  > wrote:
> Hi Todd,
> 
> I did a simple test of our ad events. We stream using Spark Streaming 
> directly into HBase, and the Data Analysts/Scientists do some 
> insight/discovery work plus some reports generation. For the reports, we use 
> SQL, and the more deeper stuff, we use Spark. In Spark, our main data 
> currency store of choice is DataFrames.
> 
> The schema is around 83 columns wide where most are of the string data type.
> 
> "event_type", "timestamp", "event_valid", "event_subtype", "user_ip", 
> "user_id", "mappable_id",
> "cookie_status", "profile_status", "user_status", "previous_timestamp", 
> "user_agent", "referer",
> "host_domain", "uri", "request_elapsed", "browser_languages", "acamp_id", 
> "creative_id",
> "location_id", “pcamp_id",
> "pdomain_id", "continent_code", "country", "region", "dma", "city", "zip", 
> "isp", "line_speed",
> "gender", "year_of_birth", "behaviors_read", "behaviors_written", 
> "key_value_pairs", "acamp_candidates",
> "tag_format", "optimizer_name", "optimizer_version", "optimizer_ip", 
> "pixel_id", “video_id",
> "video_network_id", "video_time_watched", "video_percentage_watched", 
> "video_media_type",
> "video_player_iframed", "video_player_in_view", "video_player_width", 
> "video_player_height",
> "conversion_valid_sale", "conversion_sale_amount", 
> "conversion_commission_amount", "conversion_step",
> "conversion_currency", "conversion_attribution", "conversion_offer_id", 
> "custom_info", "frequency",
> "recency_seconds", "cost", "revenue", “optimizer_acamp_id",
> "optimizer_creative_id", "optimizer_ecpm", "impression_id", "diagnostic_data",
> "user_profile_mapping_source", "latitude", "longitude", "area_code", 
> "gmt_offset", "in_dst",
> "proxy_type", "mobile_carrier", "pop", "hostname", "profile_expires", 
> "timestamp_iso", "reference_id",
> "identity_organization", "identity_method"
> 
> Most queries are like counts of how many users use what browser, how many are 
> unique users, etc. The part that scares most users is when it comes to 
> joining this data with other dimension/3rd party events tables because of 
> shear size of it.
> 
> We do what most companies do, similar to what I saw in earlier presentations 
> of Kudu. We dump data out of HBase into partitioned Parquet tables to make 
> query performance manageable.
> 
> I will coordinate with a data scientist today to do some tests. He is working 
> on identity matching/record linking of users from 2 domains: US and 
> Singapore, using probabilistic deduping algorithms. I will load the data from 
> ad events from both countries, and let him run his process against this data 
> in Kudu. I hope this will “wow” the team.
> 
> Thanks,
> Ben
> 
>> On Jun 15, 2016, at 12:47 AM, Todd Lipcon > > wrote:
>> 
>> Hi Benjamin,
>> 
>> What workload are you using for benchmarks? Using spark or something more 
>> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and 
>> some queries
>> 
>> Todd
>> 
>> Todd
>> 
>> On Jun 15, 2016 8:10 AM, "Benjamin Kim" > > wrote:
>> Hi Todd,
>> 
>> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am impressed. 
>> Compared to HBase, read and write performance are better. Write performance 
>> has the greatest improvement (> 4x), while read is > 1.5x. Albeit, these are 
>> only preliminary tests. Do you know of a way to really do some conclusive 
>> tests? I want to see if I can match your results on my 50 node cluster.
>> 
>> Thanks,
>> Ben
>> 
>>> On May 30, 2016, at 10:33 AM, Todd Lipcon >> > wrote:
>>> 
>>> On Sat, May 28, 2016 

Re: Performance Question

[Todd Lipcon]

Awesome use case. One thing to keep in mind is that spark parallelism will
be limited by the number of tablets. So, you might want to split into 10 or
so buckets per node to get the best query throughput.

Usually if you run top on some machines while running the query you can see
if it is fully utilizing the cores.

Another known issue right now is that spark locality isn't working properly
on replicated tables so you will use a lot of network traffic. For a perf
test you might want to try a table with replication count 1
On Jun 15, 2016 5:26 PM, "Benjamin Kim"  wrote:

Hi Todd,

I did a simple test of our ad events. We stream using Spark Streaming
directly into HBase, and the Data Analysts/Scientists do some
insight/discovery work plus some reports generation. For the reports, we
use SQL, and the more deeper stuff, we use Spark. In Spark, our main data
currency store of choice is DataFrames.

The schema is around 83 columns wide where most are of the string data type.

"event_type", "timestamp", "event_valid", "event_subtype", "user_ip",
"user_id", "mappable_id",
"cookie_status", "profile_status", "user_status", "previous_timestamp",
"user_agent", "referer",
"host_domain", "uri", "request_elapsed", "browser_languages", "acamp_id",
"creative_id",
"location_id", “pcamp_id",
"pdomain_id", "continent_code", "country", "region", "dma", "city", "zip",
"isp", "line_speed",
"gender", "year_of_birth", "behaviors_read", "behaviors_written",
"key_value_pairs", "acamp_candidates",
"tag_format", "optimizer_name", "optimizer_version", "optimizer_ip",
"pixel_id", “video_id",
"video_network_id", "video_time_watched", "video_percentage_watched",
"video_media_type",
"video_player_iframed", "video_player_in_view", "video_player_width",
"video_player_height",
"conversion_valid_sale", "conversion_sale_amount",
"conversion_commission_amount", "conversion_step",
"conversion_currency", "conversion_attribution", "conversion_offer_id",
"custom_info", "frequency",
"recency_seconds", "cost", "revenue", “optimizer_acamp_id",
"optimizer_creative_id", "optimizer_ecpm", "impression_id",
"diagnostic_data",
"user_profile_mapping_source", "latitude", "longitude", "area_code",
"gmt_offset", "in_dst",
"proxy_type", "mobile_carrier", "pop", "hostname", "profile_expires",
"timestamp_iso", "reference_id",
"identity_organization", "identity_method"

Most queries are like counts of how many users use what browser, how many
are unique users, etc. The part that scares most users is when it comes to
joining this data with other dimension/3rd party events tables because of
shear size of it.

We do what most companies do, similar to what I saw in earlier
presentations of Kudu. We dump data out of HBase into partitioned Parquet
tables to make query performance manageable.

I will coordinate with a data scientist today to do some tests. He is
working on identity matching/record linking of users from 2 domains: US and
Singapore, using probabilistic deduping algorithms. I will load the data
from ad events from both countries, and let him run his process against
this data in Kudu. I hope this will “wow” the team.

Thanks,
Ben

On Jun 15, 2016, at 12:47 AM, Todd Lipcon  wrote:

Hi Benjamin,

What workload are you using for benchmarks? Using spark or something more
custom? rdd or data frame or SQL, etc? Maybe you can share the schema and
some queries

Todd

Todd
On Jun 15, 2016 8:10 AM, "Benjamin Kim"  wrote:

> Hi Todd,
>
> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am
> impressed. Compared to HBase, read and write performance are better. Write
> performance has the greatest improvement (> 4x), while read is > 1.5x.
> Albeit, these are only preliminary tests. Do you know of a way to really do
> some conclusive tests? I want to see if I can match your results on my 50
> node cluster.
>
> Thanks,
> Ben
>
> On May 30, 2016, at 10:33 AM, Todd Lipcon  wrote:
>
> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim  wrote:
>
>> Todd,
>>
>> It sounds like Kudu can possibly top or match those numbers put out by
>> Aerospike. Do you have any performance statistics published or any
>> instructions as to measure them myself as good way to test? In addition,
>> this will be a test using Spark, so should I wait for Kudu version 0.9.0
>> where support will be built in?
>>
>
> We don't have a lot of benchmarks published yet, especially on the write
> side. I've found that thorough cross-system benchmarks are very difficult
> to do fairly and accurately, and often times users end up misguided if they
> pay too much attention to them :) So, given a finite number of developers
> working on Kudu, I think we've tended to spend more time on the project
> itself and less time focusing on "competition". I'm sure there are use
> cases where Kudu will beat out Aerospike, and probably use cases where
> Aerospike will beat Kudu as well.
>
> From my perspective, it would be 

Re: Performance Question

[Benjamin Kim]

Hi Todd,

I did a simple test of our ad events. We stream using Spark Streaming directly 
into HBase, and the Data Analysts/Scientists do some insight/discovery work 
plus some reports generation. For the reports, we use SQL, and the more deeper 
stuff, we use Spark. In Spark, our main data currency store of choice is 
DataFrames.

The schema is around 83 columns wide where most are of the string data type.

"event_type", "timestamp", "event_valid", "event_subtype", "user_ip", 
"user_id", "mappable_id",
"cookie_status", "profile_status", "user_status", "previous_timestamp", 
"user_agent", "referer",
"host_domain", "uri", "request_elapsed", "browser_languages", "acamp_id", 
"creative_id",
"location_id", “pcamp_id",
"pdomain_id", "continent_code", "country", "region", "dma", "city", "zip", 
"isp", "line_speed",
"gender", "year_of_birth", "behaviors_read", "behaviors_written", 
"key_value_pairs", "acamp_candidates",
"tag_format", "optimizer_name", "optimizer_version", "optimizer_ip", 
"pixel_id", “video_id",
"video_network_id", "video_time_watched", "video_percentage_watched", 
"video_media_type",
"video_player_iframed", "video_player_in_view", "video_player_width", 
"video_player_height",
"conversion_valid_sale", "conversion_sale_amount", 
"conversion_commission_amount", "conversion_step",
"conversion_currency", "conversion_attribution", "conversion_offer_id", 
"custom_info", "frequency",
"recency_seconds", "cost", "revenue", “optimizer_acamp_id",
"optimizer_creative_id", "optimizer_ecpm", "impression_id", "diagnostic_data",
"user_profile_mapping_source", "latitude", "longitude", "area_code", 
"gmt_offset", "in_dst",
"proxy_type", "mobile_carrier", "pop", "hostname", "profile_expires", 
"timestamp_iso", "reference_id",
"identity_organization", "identity_method"

Most queries are like counts of how many users use what browser, how many are 
unique users, etc. The part that scares most users is when it comes to joining 
this data with other dimension/3rd party events tables because of shear size of 
it.

We do what most companies do, similar to what I saw in earlier presentations of 
Kudu. We dump data out of HBase into partitioned Parquet tables to make query 
performance manageable.

I will coordinate with a data scientist today to do some tests. He is working 
on identity matching/record linking of users from 2 domains: US and Singapore, 
using probabilistic deduping algorithms. I will load the data from ad events 
from both countries, and let him run his process against this data in Kudu. I 
hope this will “wow” the team.

Thanks,
Ben

> On Jun 15, 2016, at 12:47 AM, Todd Lipcon  wrote:
> 
> Hi Benjamin,
> 
> What workload are you using for benchmarks? Using spark or something more 
> custom? rdd or data frame or SQL, etc? Maybe you can share the schema and 
> some queries
> 
> Todd
> 
> Todd
> 
> On Jun 15, 2016 8:10 AM, "Benjamin Kim"  > wrote:
> Hi Todd,
> 
> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am impressed. 
> Compared to HBase, read and write performance are better. Write performance 
> has the greatest improvement (> 4x), while read is > 1.5x. Albeit, these are 
> only preliminary tests. Do you know of a way to really do some conclusive 
> tests? I want to see if I can match your results on my 50 node cluster.
> 
> Thanks,
> Ben
> 
>> On May 30, 2016, at 10:33 AM, Todd Lipcon > > wrote:
>> 
>> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim > > wrote:
>> Todd,
>> 
>> It sounds like Kudu can possibly top or match those numbers put out by 
>> Aerospike. Do you have any performance statistics published or any 
>> instructions as to measure them myself as good way to test? In addition, 
>> this will be a test using Spark, so should I wait for Kudu version 0.9.0 
>> where support will be built in?
>> 
>> We don't have a lot of benchmarks published yet, especially on the write 
>> side. I've found that thorough cross-system benchmarks are very difficult to 
>> do fairly and accurately, and often times users end up misguided if they pay 
>> too much attention to them :) So, given a finite number of developers 
>> working on Kudu, I think we've tended to spend more time on the project 
>> itself and less time focusing on "competition". I'm sure there are use cases 
>> where Kudu will beat out Aerospike, and probably use cases where Aerospike 
>> will beat Kudu as well.
>> 
>> From my perspective, it would be great if you can share some details of your 
>> workload, especially if there are some areas you're finding Kudu lacking. 
>> Maybe we can spot some easy code changes we could make to improve 
>> performance, or suggest a tuning variable you could change.
>> 
>> -Todd
>> 
>> 
>>> On May 27, 2016, at 9:19 PM, Todd Lipcon >> > wrote:
>>> 
>>> On Fri, May 27, 2016 at 8:20 

Re: Performance Question

[Todd Lipcon]

Hi Benjamin,

What workload are you using for benchmarks? Using spark or something more
custom? rdd or data frame or SQL, etc? Maybe you can share the schema and
some queries

Todd

Todd
On Jun 15, 2016 8:10 AM, "Benjamin Kim"  wrote:

> Hi Todd,
>
> Now that Kudu 0.9.0 is out. I have done some tests. Already, I am
> impressed. Compared to HBase, read and write performance are better. Write
> performance has the greatest improvement (> 4x), while read is > 1.5x.
> Albeit, these are only preliminary tests. Do you know of a way to really do
> some conclusive tests? I want to see if I can match your results on my 50
> node cluster.
>
> Thanks,
> Ben
>
> On May 30, 2016, at 10:33 AM, Todd Lipcon  wrote:
>
> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim  wrote:
>
>> Todd,
>>
>> It sounds like Kudu can possibly top or match those numbers put out by
>> Aerospike. Do you have any performance statistics published or any
>> instructions as to measure them myself as good way to test? In addition,
>> this will be a test using Spark, so should I wait for Kudu version 0.9.0
>> where support will be built in?
>>
>
> We don't have a lot of benchmarks published yet, especially on the write
> side. I've found that thorough cross-system benchmarks are very difficult
> to do fairly and accurately, and often times users end up misguided if they
> pay too much attention to them :) So, given a finite number of developers
> working on Kudu, I think we've tended to spend more time on the project
> itself and less time focusing on "competition". I'm sure there are use
> cases where Kudu will beat out Aerospike, and probably use cases where
> Aerospike will beat Kudu as well.
>
> From my perspective, it would be great if you can share some details of
> your workload, especially if there are some areas you're finding Kudu
> lacking. Maybe we can spot some easy code changes we could make to improve
> performance, or suggest a tuning variable you could change.
>
> -Todd
>
>
>> On May 27, 2016, at 9:19 PM, Todd Lipcon  wrote:
>>
>> On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim  wrote:
>>
>>> Hi Mike,
>>>
>>> First of all, thanks for the link. It looks like an interesting read. I
>>> checked that Aerospike is currently at version 3.8.2.3, and in the article,
>>> they are evaluating version 3.5.4. The main thing that impressed me was
>>> their claim that they can beat Cassandra and HBase by 8x for writing and
>>> 25x for reading. Their big claim to fame is that Aerospike can write 1M
>>> records per second with only 50 nodes. I wanted to see if this is real.
>>>
>>
>> 1M records per second on 50 nodes is pretty doable by Kudu as well,
>> depending on the size of your records and the insertion order. I've been
>> playing with a ~70 node cluster recently and seen 1M+ writes/second
>> sustained, and bursting above 4M. These are 1KB rows with 11 columns, and
>> with pretty old HDD-only nodes. I think newer flash-based nodes could do
>> better.
>>
>>
>>>
>>> To answer your questions, we have a DMP with user profiles with many
>>> attributes. We create segmentation information off of these attributes to
>>> classify them. Then, we can target advertising appropriately for our sales
>>> department. Much of the data processing is for applying models on all or if
>>> not most of every profile’s attributes to find similarities (nearest
>>> neighbor/clustering) over a large number of rows when batch processing or a
>>> small subset of rows for quick online scoring. So, our use case is a
>>> typical advanced analytics scenario. We have tried HBase, but it doesn’t
>>> work well for these types of analytics.
>>>
>>> I read, that Aerospike in the release notes, they did do many
>>> improvements for batch and scan operations.
>>>
>>> I wonder what your thoughts are for using Kudu for this.
>>>
>>
>> Sounds like a good Kudu use case to me. I've heard great things about
>> Aerospike for the low latency random access portion, but I've also heard
>> that it's _very_ expensive, and not particularly suited to the columnar
>> scan workload. Lastly, I think the Apache license of Kudu is much more
>> appealing than the AGPL3 used by Aerospike. But, that's not really a direct
>> answer to the performance question :)
>>
>>
>>>
>>> Thanks,
>>> Ben
>>>
>>>
>>> On May 27, 2016, at 6:21 PM, Mike Percy  wrote:
>>>
>>> Have you considered whether you have a scan heavy or a random access
>>> heavy workload? Have you considered whether you always access / update a
>>> whole row vs only a partial row? Kudu is a column store so has some
>>> awesome performance characteristics when you are doing a lot of scanning of
>>> just a couple of columns.
>>>
>>> I don't know the answer to your question but if your concern is
>>> performance then I would be interested in seeing comparisons from a perf
>>> perspective on certain workloads.
>>>
>>> Finally, a year ago 

Re: Performance Question

[Benjamin Kim]

Hi Todd,

Now that Kudu 0.9.0 is out. I have done some tests. Already, I am impressed. 
Compared to HBase, read and write performance are better. Write performance has 
the greatest improvement (> 4x), while read is > 1.5x. Albeit, these are only 
preliminary tests. Do you know of a way to really do some conclusive tests? I 
want to see if I can match your results on my 50 node cluster.

Thanks,
Ben

> On May 30, 2016, at 10:33 AM, Todd Lipcon  wrote:
> 
> On Sat, May 28, 2016 at 7:12 AM, Benjamin Kim  > wrote:
> Todd,
> 
> It sounds like Kudu can possibly top or match those numbers put out by 
> Aerospike. Do you have any performance statistics published or any 
> instructions as to measure them myself as good way to test? In addition, this 
> will be a test using Spark, so should I wait for Kudu version 0.9.0 where 
> support will be built in?
> 
> We don't have a lot of benchmarks published yet, especially on the write 
> side. I've found that thorough cross-system benchmarks are very difficult to 
> do fairly and accurately, and often times users end up misguided if they pay 
> too much attention to them :) So, given a finite number of developers working 
> on Kudu, I think we've tended to spend more time on the project itself and 
> less time focusing on "competition". I'm sure there are use cases where Kudu 
> will beat out Aerospike, and probably use cases where Aerospike will beat 
> Kudu as well.
> 
> From my perspective, it would be great if you can share some details of your 
> workload, especially if there are some areas you're finding Kudu lacking. 
> Maybe we can spot some easy code changes we could make to improve 
> performance, or suggest a tuning variable you could change.
> 
> -Todd
> 
> 
>> On May 27, 2016, at 9:19 PM, Todd Lipcon > > wrote:
>> 
>> On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim > > wrote:
>> Hi Mike,
>> 
>> First of all, thanks for the link. It looks like an interesting read. I 
>> checked that Aerospike is currently at version 3.8.2.3, and in the article, 
>> they are evaluating version 3.5.4. The main thing that impressed me was 
>> their claim that they can beat Cassandra and HBase by 8x for writing and 25x 
>> for reading. Their big claim to fame is that Aerospike can write 1M records 
>> per second with only 50 nodes. I wanted to see if this is real.
>> 
>> 1M records per second on 50 nodes is pretty doable by Kudu as well, 
>> depending on the size of your records and the insertion order. I've been 
>> playing with a ~70 node cluster recently and seen 1M+ writes/second 
>> sustained, and bursting above 4M. These are 1KB rows with 11 columns, and 
>> with pretty old HDD-only nodes. I think newer flash-based nodes could do 
>> better.
>>  
>> 
>> To answer your questions, we have a DMP with user profiles with many 
>> attributes. We create segmentation information off of these attributes to 
>> classify them. Then, we can target advertising appropriately for our sales 
>> department. Much of the data processing is for applying models on all or if 
>> not most of every profile’s attributes to find similarities (nearest 
>> neighbor/clustering) over a large number of rows when batch processing or a 
>> small subset of rows for quick online scoring. So, our use case is a typical 
>> advanced analytics scenario. We have tried HBase, but it doesn’t work well 
>> for these types of analytics.
>> 
>> I read, that Aerospike in the release notes, they did do many improvements 
>> for batch and scan operations.
>> 
>> I wonder what your thoughts are for using Kudu for this.
>> 
>> Sounds like a good Kudu use case to me. I've heard great things about 
>> Aerospike for the low latency random access portion, but I've also heard 
>> that it's _very_ expensive, and not particularly suited to the columnar scan 
>> workload. Lastly, I think the Apache license of Kudu is much more appealing 
>> than the AGPL3 used by Aerospike. But, that's not really a direct answer to 
>> the performance question :)
>>  
>> 
>> Thanks,
>> Ben
>> 
>> 
>>> On May 27, 2016, at 6:21 PM, Mike Percy >> > wrote:
>>> 
>>> Have you considered whether you have a scan heavy or a random access heavy 
>>> workload? Have you considered whether you always access / update a whole 
>>> row vs only a partial row? Kudu is a column store so has some awesome 
>>> performance characteristics when you are doing a lot of scanning of just a 
>>> couple of columns.
>>> 
>>> I don't know the answer to your question but if your concern is performance 
>>> then I would be interested in seeing comparisons from a perf perspective on 
>>> certain workloads.
>>> 
>>> Finally, a year ago Aerospike did quite poorly in a Jepsen test: 
>>> https://aphyr.com/posts/324-jepsen-aerospike 
>>> 

Re: Performance Question

[Benjamin Kim]

Todd,

It sounds like Kudu can possibly top or match those numbers put out by 
Aerospike. Do you have any performance statistics published or any instructions 
as to measure them myself as good way to test? In addition, this will be a test 
using Spark, so should I wait for Kudu version 0.9.0 where support will be 
built in?

Thanks,
Ben


> On May 27, 2016, at 9:19 PM, Todd Lipcon  wrote:
> 
> On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim  > wrote:
> Hi Mike,
> 
> First of all, thanks for the link. It looks like an interesting read. I 
> checked that Aerospike is currently at version 3.8.2.3, and in the article, 
> they are evaluating version 3.5.4. The main thing that impressed me was their 
> claim that they can beat Cassandra and HBase by 8x for writing and 25x for 
> reading. Their big claim to fame is that Aerospike can write 1M records per 
> second with only 50 nodes. I wanted to see if this is real.
> 
> 1M records per second on 50 nodes is pretty doable by Kudu as well, depending 
> on the size of your records and the insertion order. I've been playing with a 
> ~70 node cluster recently and seen 1M+ writes/second sustained, and bursting 
> above 4M. These are 1KB rows with 11 columns, and with pretty old HDD-only 
> nodes. I think newer flash-based nodes could do better.
>  
> 
> To answer your questions, we have a DMP with user profiles with many 
> attributes. We create segmentation information off of these attributes to 
> classify them. Then, we can target advertising appropriately for our sales 
> department. Much of the data processing is for applying models on all or if 
> not most of every profile’s attributes to find similarities (nearest 
> neighbor/clustering) over a large number of rows when batch processing or a 
> small subset of rows for quick online scoring. So, our use case is a typical 
> advanced analytics scenario. We have tried HBase, but it doesn’t work well 
> for these types of analytics.
> 
> I read, that Aerospike in the release notes, they did do many improvements 
> for batch and scan operations.
> 
> I wonder what your thoughts are for using Kudu for this.
> 
> Sounds like a good Kudu use case to me. I've heard great things about 
> Aerospike for the low latency random access portion, but I've also heard that 
> it's _very_ expensive, and not particularly suited to the columnar scan 
> workload. Lastly, I think the Apache license of Kudu is much more appealing 
> than the AGPL3 used by Aerospike. But, that's not really a direct answer to 
> the performance question :)
>  
> 
> Thanks,
> Ben
> 
> 
>> On May 27, 2016, at 6:21 PM, Mike Percy > > wrote:
>> 
>> Have you considered whether you have a scan heavy or a random access heavy 
>> workload? Have you considered whether you always access / update a whole row 
>> vs only a partial row? Kudu is a column store so has some awesome 
>> performance characteristics when you are doing a lot of scanning of just a 
>> couple of columns.
>> 
>> I don't know the answer to your question but if your concern is performance 
>> then I would be interested in seeing comparisons from a perf perspective on 
>> certain workloads.
>> 
>> Finally, a year ago Aerospike did quite poorly in a Jepsen test: 
>> https://aphyr.com/posts/324-jepsen-aerospike 
>> 
>> 
>> I wonder if they have addressed any of those issues.
>> 
>> Mike
>> 
>> On Friday, May 27, 2016, Benjamin Kim > > wrote:
>> I am just curious. How will Kudu compare with Aerospike 
>> (http://www.aerospike.com )? I went to a Spark 
>> Roadshow and found out about this piece of software. It appears to fit our 
>> use case perfectly since we are an ad-tech company trying to leverage our 
>> user profiles data. Plus, it already has a Spark connector and has a 
>> SQL-like client. The tables can be accessed using Spark SQL DataFrames and, 
>> also, made into SQL tables for direct use with Spark SQL ODBC/JDBC 
>> Thriftserver. I see from the work done here 
>> http://gerrit.cloudera.org:8080/#/c/2992/ 
>>  that the Spark integration is 
>> well underway and, from the looks of it lately, almost complete. I would 
>> prefer to use Kudu since we are already a Cloudera shop, and Kudu is easy to 
>> deploy and configure using Cloudera Manager. I also hope that some of 
>> Aerospike’s speed optimization techniques can make it into Kudu in the 
>> future, if they have not been already thought of or included.
>> 
>> Just some thoughts…
>> 
>> Cheers,
>> Ben
>> 
>> 
>> -- 
>> --
>> Mike Percy
>> Software Engineer, Cloudera
>> 
>> 
> 
> 
> 
> 
> -- 
> Todd Lipcon
> Software Engineer, Cloudera

Re: Performance Question

[Todd Lipcon]

On Fri, May 27, 2016 at 8:20 PM, Benjamin Kim  wrote:

> Hi Mike,
>
> First of all, thanks for the link. It looks like an interesting read. I
> checked that Aerospike is currently at version 3.8.2.3, and in the article,
> they are evaluating version 3.5.4. The main thing that impressed me was
> their claim that they can beat Cassandra and HBase by 8x for writing and
> 25x for reading. Their big claim to fame is that Aerospike can write 1M
> records per second with only 50 nodes. I wanted to see if this is real.
>

1M records per second on 50 nodes is pretty doable by Kudu as well,
depending on the size of your records and the insertion order. I've been
playing with a ~70 node cluster recently and seen 1M+ writes/second
sustained, and bursting above 4M. These are 1KB rows with 11 columns, and
with pretty old HDD-only nodes. I think newer flash-based nodes could do
better.


>
> To answer your questions, we have a DMP with user profiles with many
> attributes. We create segmentation information off of these attributes to
> classify them. Then, we can target advertising appropriately for our sales
> department. Much of the data processing is for applying models on all or if
> not most of every profile’s attributes to find similarities (nearest
> neighbor/clustering) over a large number of rows when batch processing or a
> small subset of rows for quick online scoring. So, our use case is a
> typical advanced analytics scenario. We have tried HBase, but it doesn’t
> work well for these types of analytics.
>
> I read, that Aerospike in the release notes, they did do many improvements
> for batch and scan operations.
>
> I wonder what your thoughts are for using Kudu for this.
>

Sounds like a good Kudu use case to me. I've heard great things about
Aerospike for the low latency random access portion, but I've also heard
that it's _very_ expensive, and not particularly suited to the columnar
scan workload. Lastly, I think the Apache license of Kudu is much more
appealing than the AGPL3 used by Aerospike. But, that's not really a direct
answer to the performance question :)


>
> Thanks,
> Ben
>
>
> On May 27, 2016, at 6:21 PM, Mike Percy  wrote:
>
> Have you considered whether you have a scan heavy or a random access heavy
> workload? Have you considered whether you always access / update a whole
> row vs only a partial row? Kudu is a column store so has some
> awesome performance characteristics when you are doing a lot of scanning of
> just a couple of columns.
>
> I don't know the answer to your question but if your concern is
> performance then I would be interested in seeing comparisons from a perf
> perspective on certain workloads.
>
> Finally, a year ago Aerospike did quite poorly in a Jepsen test:
> https://aphyr.com/posts/324-jepsen-aerospike
>
> I wonder if they have addressed any of those issues.
>
> Mike
>
> On Friday, May 27, 2016, Benjamin Kim  wrote:
>
>> I am just curious. How will Kudu compare with Aerospike (
>> http://www.aerospike.com)? I went to a Spark Roadshow and found out
>> about this piece of software. It appears to fit our use case perfectly
>> since we are an ad-tech company trying to leverage our user profiles data.
>> Plus, it already has a Spark connector and has a SQL-like client. The
>> tables can be accessed using Spark SQL DataFrames and, also, made into SQL
>> tables for direct use with Spark SQL ODBC/JDBC Thriftserver. I see from the
>> work done here http://gerrit.cloudera.org:8080/#/c/2992/ that the Spark
>> integration is well underway and, from the looks of it lately, almost
>> complete. I would prefer to use Kudu since we are already a Cloudera shop,
>> and Kudu is easy to deploy and configure using Cloudera Manager. I also
>> hope that some of Aerospike’s speed optimization techniques can make it
>> into Kudu in the future, if they have not been already thought of or
>> included.
>>
>> Just some thoughts…
>>
>> Cheers,
>> Ben
>
>
>
> --
> --
> Mike Percy
> Software Engineer, Cloudera
>
>
>
>


-- 
Todd Lipcon
Software Engineer, Cloudera

Re: Performance Question

[Benjamin Kim]

Hi Mike,

First of all, thanks for the link. It looks like an interesting read. I checked 
that Aerospike is currently at version 3.8.2.3, and in the article, they are 
evaluating version 3.5.4. The main thing that impressed me was their claim that 
they can beat Cassandra and HBase by 8x for writing and 25x for reading. Their 
big claim to fame is that Aerospike can write 1M records per second with only 
50 nodes. I wanted to see if this is real.

To answer your questions, we have a DMP with user profiles with many 
attributes. We create segmentation information off of these attributes to 
classify them. Then, we can target advertising appropriately for our sales 
department. Much of the data processing is for applying models on all or if not 
most of every profile’s attributes to find similarities (nearest 
neighbor/clustering) over a large number of rows when batch processing or a 
small subset of rows for quick online scoring. So, our use case is a typical 
advanced analytics scenario. We have tried HBase, but it doesn’t work well for 
these types of analytics.

I read, that Aerospike in the release notes, they did do many improvements for 
batch and scan operations.

I wonder what your thoughts are for using Kudu for this.

Thanks,
Ben


> On May 27, 2016, at 6:21 PM, Mike Percy  wrote:
> 
> Have you considered whether you have a scan heavy or a random access heavy 
> workload? Have you considered whether you always access / update a whole row 
> vs only a partial row? Kudu is a column store so has some awesome performance 
> characteristics when you are doing a lot of scanning of just a couple of 
> columns.
> 
> I don't know the answer to your question but if your concern is performance 
> then I would be interested in seeing comparisons from a perf perspective on 
> certain workloads.
> 
> Finally, a year ago Aerospike did quite poorly in a Jepsen test: 
> https://aphyr.com/posts/324-jepsen-aerospike 
> 
> 
> I wonder if they have addressed any of those issues.
> 
> Mike
> 
> On Friday, May 27, 2016, Benjamin Kim  > wrote:
> I am just curious. How will Kudu compare with Aerospike 
> (http://www.aerospike.com )? I went to a Spark 
> Roadshow and found out about this piece of software. It appears to fit our 
> use case perfectly since we are an ad-tech company trying to leverage our 
> user profiles data. Plus, it already has a Spark connector and has a SQL-like 
> client. The tables can be accessed using Spark SQL DataFrames and, also, made 
> into SQL tables for direct use with Spark SQL ODBC/JDBC Thriftserver. I see 
> from the work done here http://gerrit.cloudera.org:8080/#/c/2992/ 
>  that the Spark integration is 
> well underway and, from the looks of it lately, almost complete. I would 
> prefer to use Kudu since we are already a Cloudera shop, and Kudu is easy to 
> deploy and configure using Cloudera Manager. I also hope that some of 
> Aerospike’s speed optimization techniques can make it into Kudu in the 
> future, if they have not been already thought of or included.
> 
> Just some thoughts…
> 
> Cheers,
> Ben
> 
> 
> -- 
> --
> Mike Percy
> Software Engineer, Cloudera
> 
> 

Re: Performance Question

[Mike Percy]

Have you considered whether you have a scan heavy or a random access heavy
workload? Have you considered whether you always access / update a whole
row vs only a partial row? Kudu is a column store so has some
awesome performance characteristics when you are doing a lot of scanning of
just a couple of columns.

I don't know the answer to your question but if your concern is performance
then I would be interested in seeing comparisons from a perf perspective on
certain workloads.

Finally, a year ago Aerospike did quite poorly in a Jepsen test:
https://aphyr.com/posts/324-jepsen-aerospike

I wonder if they have addressed any of those issues.

Mike

On Friday, May 27, 2016, Benjamin Kim  wrote:

> I am just curious. How will Kudu compare with Aerospike (
> http://www.aerospike.com)? I went to a Spark Roadshow and found out about
> this piece of software. It appears to fit our use case perfectly since we
> are an ad-tech company trying to leverage our user profiles data. Plus, it
> already has a Spark connector and has a SQL-like client. The tables can be
> accessed using Spark SQL DataFrames and, also, made into SQL tables for
> direct use with Spark SQL ODBC/JDBC Thriftserver. I see from the work done
> here http://gerrit.cloudera.org:8080/#/c/2992/ that the Spark integration
> is well underway and, from the looks of it lately, almost complete. I would
> prefer to use Kudu since we are already a Cloudera shop, and Kudu is easy
> to deploy and configure using Cloudera Manager. I also hope that some of
> Aerospike’s speed optimization techniques can make it into Kudu in the
> future, if they have not been already thought of or included.
>
> Just some thoughts…
>
> Cheers,
> Ben



-- 
--
Mike Percy
Software Engineer, Cloudera