phoenix-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Josh Elser <els...@apache.org>
Subject Re: Missing content in phoenix after writing from Spark
Date Mon, 17 Sep 2018 23:45:50 GMT
As I said earlier, the expectation is that you use the 
phoenix-client.jar and phoenix-spark2.jar for the phoenix-spark 
integration with spark2.

You do not need to reference all of these jars by hand. We create the 
jars with all of the necessary dependencies bundled to specifically 
avoid creating this problem for users.

On 9/17/18 3:27 PM, Saif Addin wrote:
> Thanks for the patience, sorry maybe I sent incomplete information. We 
> are loading the following jars and still getting: */executor 1): 
> java.lang.NoClassDefFoundError: Could not initialize class 
> org.apache.phoenix.query.QueryServicesOptions/*
> */
> /*
> http://central.maven.org/maven2/org/apache/hbase/hbase-client/2.1.0/hbase-client-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-common/2.1.0/hbase-common-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-hadoop-compat/2.1.0/hbase-hadoop-compat-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-mapreduce/2.1.0/hbase-mapreduce-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/thirdparty/hbase-shaded-miscellaneous/2.1.0/hbase-shaded-miscellaneous-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-protocol/2.1.0/hbase-protocol-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-protocol-shaded/2.1.0/hbase-protocol-shaded-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/thirdparty/hbase-shaded-protobuf/2.1.0/hbase-shaded-protobuf-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/thirdparty/hbase-shaded-netty/2.1.0/hbase-shaded-netty-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-server/2.1.0/hbase-server-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-hadoop2-compat/2.1.0/hbase-hadoop2-compat-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-metrics/2.1.0/hbase-metrics-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-metrics-api/2.1.0/hbase-metrics-api-2.1.0.jar
> http://central.maven.org/maven2/org/apache/hbase/hbase-zookeeper/2.1.0/hbase-zookeeper-2.1.0.jar
> 
> http://central.maven.org/maven2/org/apache/phoenix/phoenix-spark/5.0.0-HBase-2.0/phoenix-spark-5.0.0-HBase-2.0.jar
> http://central.maven.org/maven2/org/apache/phoenix/phoenix-core/5.0.0-HBase-2.0/phoenix-core-5.0.0-HBase-2.0.jar
> http://central.maven.org/maven2/org/apache/phoenix/phoenix-queryserver/5.0.0-HBase-2.0/phoenix-queryserver-5.0.0-HBase-2.0.jar
> http://central.maven.org/maven2/org/apache/phoenix/phoenix-queryserver-client/5.0.0-HBase-2.0/phoenix-queryserver-client-5.0.0-HBase-2.0.jar
> 
> http://central.maven.org/maven2/org/apache/twill/twill-zookeeper/0.13.0/twill-zookeeper-0.13.0.jar
> http://central.maven.org/maven2/org/apache/twill/twill-discovery-core/0.13.0/twill-discovery-core-0.13.0.jar
> 
> Not sure which one I could be missing??
> 
> On Fri, Sep 14, 2018 at 7:34 PM Josh Elser <elserj@apache.org 
> <mailto:elserj@apache.org>> wrote:
> 
>     Uh, you're definitely not using the right JARs :)
> 
>     You'll want the phoenix-client.jar for the Phoenix JDBC driver and the
>     phoenix-spark.jar for the Phoenix RDD.
> 
>     On 9/14/18 1:08 PM, Saif Addin wrote:
>      > Hi, I am attempting to make connection with Spark but no success
>     so far.
>      >
>      > For writing into Phoenix, I am trying this:
>      >
>      > tdd.toDF("ID", "COL1", "COL2",
>      > "COL3").write.format("org.apache.phoenix.spark").option("zkUrl",
>      > "zookeper-host-url:2181").option("table",
>      > htablename).mode("overwrite").save()
>      >
>      > But getting:
>      > *java.sql.SQLException: ERROR 103 (08004): Unable to establish
>     connection.*
>      > *
>      > *
>      > For reading, on the other hand, attempting this:
>      >
>      > val hbConf = HBaseConfiguration.create()
>      > val hbaseSitePath = "/etc/hbase/conf/hbase-site.xml"
>      > hbConf.addResource(new Path(hbaseSitePath))
>      >
>      > spark.sqlContext.phoenixTableAsDataFrame("VISTA_409X68",
>     Array("ID"),
>      > conf = hbConf)
>      >
>      > Gets me
>      > *java.lang.NoClassDefFoundError: Could not initialize class
>      > org.apache.phoenix.query.QueryServicesOptions*
>      > *
>      > *
>      > I have added phoenix-queryserver-5.0.0-HBase-2.0.jar and
>      > phoenix-queryserver-client-5.0.0-HBase-2.0.jar
>      > Any thoughts? I have an hbase-site.xml file with more
>     configuration but
>      > not sure how to get it to be read in the saving instance.
>      > Thanks
>      >
>      > On Thu, Sep 13, 2018 at 11:38 AM Josh Elser <elserj@apache.org
>     <mailto:elserj@apache.org>
>      > <mailto:elserj@apache.org <mailto:elserj@apache.org>>> wrote:
>      >
>      >     Pretty sure we ran tests with Spark 2.3 with Phoenix 5.0. Not
>     sure if
>      >     Spark has already moved beyond that.
>      >
>      >     On 9/12/18 11:00 PM, Saif Addin wrote:
>      >      > Thanks, we'll try Spark Connector then. Thought it didn't
>     support
>      >     newest
>      >      > Spark Versions
>      >      >
>      >      > On Wed, Sep 12, 2018 at 11:03 PM Jaanai Zhang
>      >     <cloud.poster@gmail.com <mailto:cloud.poster@gmail.com>
>     <mailto:cloud.poster@gmail.com <mailto:cloud.poster@gmail.com>>
>      >      > <mailto:cloud.poster@gmail.com
>     <mailto:cloud.poster@gmail.com> <mailto:cloud.poster@gmail.com
>     <mailto:cloud.poster@gmail.com>>>>
>      >     wrote:
>      >      >
>      >      >     It seems columns data missing mapping information of the
>      >     schema. if
>      >      >     you want to use this way to write HBase table,  you
>     can create an
>      >      >     HBase table and uses Phoenix mapping it.
>      >      >
>      >      >     ----------------------------------------
>      >      >         Jaanai Zhang
>      >      >         Best regards!
>      >      >
>      >      >
>      >      >
>      >      >     Thomas D'Silva <tdsilva@salesforce.com
>     <mailto:tdsilva@salesforce.com>
>      >     <mailto:tdsilva@salesforce.com <mailto:tdsilva@salesforce.com>>
>      >      >     <mailto:tdsilva@salesforce.com
>     <mailto:tdsilva@salesforce.com>
>      >     <mailto:tdsilva@salesforce.com
>     <mailto:tdsilva@salesforce.com>>>> 于2018年9月13日周四 上午6:03写道:
>      >      >
>      >      >         Is there a reason you didn't use the
>     spark-connector to
>      >      >         serialize your data?
>      >      >
>      >      >         On Wed, Sep 12, 2018 at 2:28 PM, Saif Addin
>      >     <saif1988@gmail.com <mailto:saif1988@gmail.com>
>     <mailto:saif1988@gmail.com <mailto:saif1988@gmail.com>>
>      >      >         <mailto:saif1988@gmail.com
>     <mailto:saif1988@gmail.com> <mailto:saif1988@gmail.com
>     <mailto:saif1988@gmail.com>>>>
>      >     wrote:
>      >      >
>      >      >             Thank you Josh! That was helpful. Indeed,
>     there was a
>      >     salt
>      >      >             bucket on the table, and the key-column now
shows
>      >     correctly.
>      >      >
>      >      >             However, the problem still persists in that
>     the rest
>      >     of the
>      >      >             columns show as completely empty on Phoenix
>     (appear
>      >      >             correctly on Hbase). We'll be looking into
>     this but
>      >     if you
>      >      >             have any further advice, appreciated.
>      >      >
>      >      >             Saif
>      >      >
>      >      >             On Wed, Sep 12, 2018 at 5:50 PM Josh Elser
>      >      >             <elserj@apache.org <mailto:elserj@apache.org>
>     <mailto:elserj@apache.org <mailto:elserj@apache.org>>
>      >     <mailto:elserj@apache.org <mailto:elserj@apache.org>
>     <mailto:elserj@apache.org <mailto:elserj@apache.org>>>> wrote:
>      >      >
>      >      >                 Reminder: Using Phoenix internals forces
>     you to
>      >      >                 understand exactly how
>      >      >                 the version of Phoenix that you're using
>     serializes
>      >      >                 data. Is there a
>      >      >                 reason you're not using SQL to interact
>     with Phoenix?
>      >      >
>      >      >                 Sounds to me that Phoenix is expecting
>     more data
>      >     at the
>      >      >                 head of your
>      >      >                 rowkey. Maybe a salt bucket that you've
>     defined
>      >     on the
>      >      >                 table but not
>      >      >                 created?
>      >      >
>      >      >                 On 9/12/18 4:32 PM, Saif Addin wrote:
>      >      >                  > Hi all,
>      >      >                  >
>      >      >                  > We're trying to write tables with
all
>     string
>      >     columns
>      >      >                 from spark.
>      >      >                  > We are not using the Spark Connector,
>     instead
>      >     we are
>      >      >                 directly writing
>      >      >                  > byte arrays from RDDs.
>      >      >                  >
>      >      >                  > The process works fine, and Hbase
>     receives the
>      >     data
>      >      >                 correctly, and
>      >      >                  > content is consistent.
>      >      >                  >
>      >      >                  > However reading the table from
Phoenix, we
>      >     notice the
>      >      >                 first character of
>      >      >                  > strings are missing. This sounds
like
>     it's a byte
>      >      >                 encoding issue, but
>      >      >                  > we're at loss. We're using PVarchar
to
>      >     generate bytes.
>      >      >                  >
>      >      >                  > Here's the snippet of code creating
the
>     RDD:
>      >      >                  >
>      >      >                  > val tdd = pdd.flatMap(x => {
>      >      >                  >    val rowKey =
>     PVarchar.INSTANCE.toBytes(x._1)
>      >      >                  >    for(i <- 0 until cols.length)
yield {
>      >      >                  >      other stuff for other columns
...
>      >      >                  >      ...
>      >      >                  >      (rowKey, (column1, column2,
column3))
>      >      >                  >    }
>      >      >                  > })
>      >      >                  >
>      >      >                  > ...
>      >      >                  >
>      >      >                  > We then create the following output
to
>     be written
>      >      >                 down in Hbase
>      >      >                  >
>      >      >                  > val output = tdd.map(x => {
>      >      >                  >      val rowKeyByte: Array[Byte]
= x._1
>      >      >                  >      val immutableRowKey = new
>      >      >                 ImmutableBytesWritable(rowKeyByte)
>      >      >                  >
>      >      >                  >      val kv = new KeyValue(rowKeyByte,
>      >      >                  >         
>     PVarchar.INSTANCE.toBytes(column1),
>      >      >                  >         
>     PVarchar.INSTANCE.toBytes(column2),
>      >      >                  >        PVarchar.INSTANCE.toBytes(column3)
>      >      >                  >      )
>      >      >                  >      (immutableRowKey, kv)
>      >      >                  > })
>      >      >                  >
>      >      >                  > By the way, we are using
>     *KryoSerializer* in
>      >     order to
>      >      >                 be able to
>      >      >                  > serialize all classes necessary
for Hbase
>      >     (KeyValue,
>      >      >                 BytesWritable, etc).
>      >      >                  >
>      >      >                  > The key of this table is the one
>     missing data when
>      >      >                 queried from Phoenix.
>      >      >                  > So we guess something is wrong
with the
>     byte ser.
>      >      >                  >
>      >      >                  > Any ideas? Appreciated!
>      >      >                  > Saif
>      >      >
>      >      >
>      >
> 

Mime
View raw message