DDevelopment & Design

Distributed scans with HBase

HBase is by design a columnar store, that is optimized for random reads. You just ask for a row using rowId as an identifier and you get your data instantaneously. Performing a scan on part or whole table is a completely different thing. First of all, it is sequential. Meaning it is rather slow, because it doesn’t use all the RegionServers at the same time. It is implemented that way to realize the contract of Scan command – which has to return results sorted by key. So, how to do this efficiently?HBase is by design a columnar store, that is optimized for random reads. You just ask for a row using rowId as an identifier and you get your data instantaneously. Performing a scan on part or whole table is a completely different thing. First of all, it is sequential. Meaning it is rather slow, because it doesn’t use all the RegionServers at the same time. It is implemented that way to realize the contract of Scan command – which has to return results sorted by key. So, how to do this efficiently?

HBase is by design a columnar store, that is optimized for random reads.
You just ask for a row using rowId as an identifier and you get your
data instantaneously.

Performing a scan on part or whole table is a completely different thing.
First of all, it is sequential. Meaning it is rather slow, because it
doesn’t use all the RegionServers at the same time. It is implemented
that way to realize the contract of Scan command – which has to return
results sorted by key.

So, how to do this efficiently?

The usual way of getting data from HBase is with the help of its API,
mainly Scan objects. To accomplish the task I’ll use just them. I’ll
specify startRow and stopRow, so that each Scan request will be looking
through only part of the key space.

It is crucial to note, that this whole solution works because of key
sorting properties in HBase. So, HBase scans a table according to ascending key
values. Since keys are of String type, key with value “1” is smaller
than “2”, because they are sorted lexicographicly. So, also key with value “12345” is smaller than “2”. I’ve
leveraged this property so that I’ve partitioned my whole key space according to
the first character of the key. In my case keys contain only digits. So I
have 10 ranges:

null-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-null

The speedup comes from the fact, that each range resides in its own
partition. That’s right, I’ve presplit the table to have 10 partitions.
This corresponds rather nicely with my cluster’s setup, because I have
more than 10 RegionServers. So every partition should be on different
RegionServer. It will allow the code to do the requested scan operations
in parallel – giving me this exact performance boost.

How I’ve created the input table:

$ create 'tariff_changes', { NAME => 'cf', SPLITS_FILE => 'splits.txt', VERSIONS => 50, MAX_FILESIZE => 1073741824 }

$ alter 'tariff_changes', { NAME => 'cf', SPLITS_FILE => 'splits.txt', VERSIONS => 50, MAX_FILESIZE => 1073741824 }

Split file is just something along this lines:

1
2
3
4
5
6
7
8
9
0

This tells HBase what are the rowKeys starting and ending each of the
partitions.

Ok, so after this rather lengthy introduction, what the actual code
does? It just spins of a few threads – one for each partition – and runs
a Scan request tailored to that partitions key space. This way, I got a
10x speedup for this particular scan. The execution time went down from
30 minutes to 3 for my use case.

I’ve created an example implementation of this idea. You can find it on
GitHub:
https://github.com/zygm0nt/hbase-distributed-search.

Any ideas on how to speed things up even more with HBase?

— Previous article

Simple HBase ORM

Next article —

Sonar Gerrit Plugin Release

You May Also Like
DDevelopment & Design

Thought static method can’t be easy to mock, stub nor track? Wrong!

No matter why, no matter is it a good idea. Sometimes one just wants to check or it's necessary to be done. Mock a static method, woot? Impossibru!

In pure Java world it is still a struggle. But Groovy allows you to do that really simple. Well, not groovy alone, but with a great support of Spock.

Lets move on straight to the example. To catch some context we have an abstract for the example needs. A marketing project with a set of offers. One to many.

import spock.lang.Specification

class OfferFacadeSpec extends Specification {

    OfferFacade facade = new OfferFacade()

    def setup() {
        GroovyMock(Project, global: true)
    }

    def 'delegates an add offer call to the domain with proper params'() {
        given:
            Map params = [projId: projectId, name: offerName]

        when:
            Offer returnedOffer = facade.add(params)

        then:
            1 * Project.addOffer(projectId, _) >> { projId, offer -> offer }
            returnedOffer.name == params.name

        where:
            projectId | offerName
            1         | 'an Offer'
            15        | 'whasup!?'
            123       | 'doskonała oferta - kup teraz!'
    }
}
So we test a facade responsible for handling "add offer to the project" call triggered  somewhere in a GUI.
We want to ensure that static method Project.addOffer(long, Offer) will receive correct params when java.util.Map with user form input comes to the facade.add(params).
This is unit test, so how Project.addOffer() works is out of scope. Thus we want to stub it.

The most important is a GroovyMock(Project, global: true) statement.
What it does is modifing Project class to behave like a Spock's mock. 
GroovyMock() itself is a method inherited from Specification.  The global flag is necessary to enable mocking static methods.
However when one comes to the need of mocking static method, author of Spock Framework advice to consider redesigning of implementation. It's not a bad advice, I must say.

Another important thing are assertions at then: block. First one checks an interaction, if the Project.addOffer() method was called exactly once, with a 1st argument equal to the projectId and some other param (we don't have an object instance yet to assert anything about it).
Right shit operator leads us to the stub which replaces original method implementation by such statement.
As a good stub it does nothing. The original method definition has return type Offer. The stub needs to do the same. So an offer passed as the 2nd argument is just returned.
Thanks to this we can assert about name property if it's equal with the value from params. If no return was designed the name could be checked inside the stub Closure, prefixed with an assert keyword.

Worth of  mentioning is that if you want to track interactions of original static method implementation without replacing it, then you should try using GroovySpy instead of GroovyMock.

Unfortunately static methods declared at Java object can't be treated in such ways. Though regular mocks and whole goodness of Spock can be used to test pure Java code, which is awesome anyway :)No matter why, no matter is it a good idea. Sometimes one just wants to check or it's necessary to be done. Mock a static method, woot? Impossibru!

In pure Java world it is still a struggle. But Groovy allows you to do that really simple. Well, not groovy alone, but with a great support of Spock.

Lets move on straight to the example. To catch some context we have an abstract for the example needs. A marketing project with a set of offers. One to many.

import spock.lang.Specification

class OfferFacadeSpec extends Specification {

    OfferFacade facade = new OfferFacade()

    def setup() {
        GroovyMock(Project, global: true)
    }

    def 'delegates an add offer call to the domain with proper params'() {
        given:
            Map params = [projId: projectId, name: offerName]

        when:
            Offer returnedOffer = facade.add(params)

        then:
            1 * Project.addOffer(projectId, _) >> { projId, offer -> offer }
            returnedOffer.name == params.name

        where:
            projectId | offerName
            1         | 'an Offer'
            15        | 'whasup!?'
            123       | 'doskonała oferta - kup teraz!'
    }
}
So we test a facade responsible for handling "add offer to the project" call triggered  somewhere in a GUI.
We want to ensure that static method Project.addOffer(long, Offer) will receive correct params when java.util.Map with user form input comes to the facade.add(params).
This is unit test, so how Project.addOffer() works is out of scope. Thus we want to stub it.

The most important is a GroovyMock(Project, global: true) statement.
What it does is modifing Project class to behave like a Spock's mock. 
GroovyMock() itself is a method inherited from Specification.  The global flag is necessary to enable mocking static methods.
However when one comes to the need of mocking static method, author of Spock Framework advice to consider redesigning of implementation. It's not a bad advice, I must say.

Another important thing are assertions at then: block. First one checks an interaction, if the Project.addOffer() method was called exactly once, with a 1st argument equal to the projectId and some other param (we don't have an object instance yet to assert anything about it).
Right shit operator leads us to the stub which replaces original method implementation by such statement.
As a good stub it does nothing. The original method definition has return type Offer. The stub needs to do the same. So an offer passed as the 2nd argument is just returned.
Thanks to this we can assert about name property if it's equal with the value from params. If no return was designed the name could be checked inside the stub Closure, prefixed with an assert keyword.

Worth of  mentioning is that if you want to track interactions of original static method implementation without replacing it, then you should try using GroovySpy instead of GroovyMock.

Unfortunately static methods declared at Java object can't be treated in such ways. Though regular mocks and whole goodness of Spock can be used to test pure Java code, which is awesome anyway :)