RE: Performance degradation on query analysis
Thanks Josh, you are right, we have actually disabled automatic major compaction. Now we added SYSTEM.STATS to weekly compaction and I hope this resolve the issue. -Original Message- From: Josh Elser [mailto:els...@apache.org] Sent: Tuesday, September 24, 2019 6:39 PM To: user@phoenix.apache.org Subject: Re: Performance degradation on query analysis Did you change your configuration to prevent compactions from regularly happening, Stepan? By default, you should have a major compaction run weekly which would have fixed this for you, although minor compactions would have run automatically as well to rewrite small hfiles as you are creating new one (generating new stats). On 9/19/19 4:50 PM, Ankit Singhal wrote: > Please schedule compaction on SYSTEM.STATS table to clear the old entries. > > On Thu, Sep 19, 2019 at 1:48 PM Stepan Migunov > <mailto:stepan.migu...@firstlinesoftware.com>> wrote: > > Thanks, Josh. The problem was really related to reading the > SYSTEM.STATS > table. > There were only 8,000 rows in the table, but COUNT took more than 10 > minutes. I noticed that the storage files (34) had a total size of > 10 GB. > > DELETE FROM SYSTEM.STATS did not help - the storage files are still > 10 GB, > and COUNT took a long time. > Then I truncated the table from the hbase shell. And this fixed the > problem - after UPDATE STATS for each table, everything works fine. > > Are there any known issues with SYSTEM.STATS table? Apache Phoenix > 4.13.1 > with 15 Region Servers. > > -Original Message- > From: Josh Elser [mailto:els...@apache.org <mailto:els...@apache.org>] > Sent: Tuesday, September 17, 2019 5:16 PM > To: user@phoenix.apache.org <mailto:user@phoenix.apache.org> > Subject: Re: Performance degradation on query analysis > > Can you share the output you see from the EXPLAIN? Does it differ > between > times it's "fast" and times it's "slow"? > > Sharing the table(s) DDL statements would also help, along with the > shape > and version of your cluster (e.g. Apache Phoenix 4.14.2 with 8 > RegionServers). > > Spit-balling ideas: > > Could be reads over the SYSTEM.CATALOG table or the SYSTEM.STATS > table. > > Have you looked more coarsely at the RegionServer logs/metrics? Any > obvious > saturation issues (e.g. handlers consumed, JVM GC pauses, host CPU > saturation)? > > Turn on DEBUG log4j client side (beware of chatty ZK logging) and see > if > there's something obvious from when the EXPLAIN is slow. > > On 9/17/19 3:58 AM, Stepan Migunov wrote: > > Hi > > We have an issue with our production environment - from time to > time we > > notice a significant performance degradation for some queries. > The strange > > thing is that the EXPLAIN operator for these queries takes the > same time > > as queries execution (5 minutes or more). So, I guess, the issue is > > related to query's analysis but not data extraction. Is it > possible that > > issue is related to SYSTEM.STATS access problem? Any other ideas? > > >
RE: Performance degradation on query analysis
Thanks, Josh. The problem was really related to reading the SYSTEM.STATS table. There were only 8,000 rows in the table, but COUNT took more than 10 minutes. I noticed that the storage files (34) had a total size of 10 GB. DELETE FROM SYSTEM.STATS did not help - the storage files are still 10 GB, and COUNT took a long time. Then I truncated the table from the hbase shell. And this fixed the problem - after UPDATE STATS for each table, everything works fine. Are there any known issues with SYSTEM.STATS table? Apache Phoenix 4.13.1 with 15 Region Servers. -Original Message- From: Josh Elser [mailto:els...@apache.org] Sent: Tuesday, September 17, 2019 5:16 PM To: user@phoenix.apache.org Subject: Re: Performance degradation on query analysis Can you share the output you see from the EXPLAIN? Does it differ between times it's "fast" and times it's "slow"? Sharing the table(s) DDL statements would also help, along with the shape and version of your cluster (e.g. Apache Phoenix 4.14.2 with 8 RegionServers). Spit-balling ideas: Could be reads over the SYSTEM.CATALOG table or the SYSTEM.STATS table. Have you looked more coarsely at the RegionServer logs/metrics? Any obvious saturation issues (e.g. handlers consumed, JVM GC pauses, host CPU saturation)? Turn on DEBUG log4j client side (beware of chatty ZK logging) and see if there's something obvious from when the EXPLAIN is slow. On 9/17/19 3:58 AM, Stepan Migunov wrote: > Hi > We have an issue with our production environment - from time to time we > notice a significant performance degradation for some queries. The strange > thing is that the EXPLAIN operator for these queries takes the same time > as queries execution (5 minutes or more). So, I guess, the issue is > related to query's analysis but not data extraction. Is it possible that > issue is related to SYSTEM.STATS access problem? Any other ideas? >
Performance degradation on query analysis
Hi We have an issue with our production environment - from time to time we notice a significant performance degradation for some queries. The strange thing is that the EXPLAIN operator for these queries takes the same time as queries execution (5 minutes or more). So, I guess, the issue is related to query's analysis but not data extraction. Is it possible that issue is related to SYSTEM.STATS access problem? Any other ideas?
RE: Phoenix ODBC driver limitations
Yes, I read this. But the document says "This needs to be set at client and server both". I've been confused - what is "client" in case of ODBC-connection. I supposed that driver, but it seems queryserver. -Original Message- From: Francis Chuang [mailto:francischu...@apache.org] Sent: Thursday, May 24, 2018 1:35 AM To: user@phoenix.apache.org Subject: Re: Phoenix ODBC driver limitations Namespace mapping is something you need to enable on the server (it's off by default). See documentation for enabling it here: http://phoenix.apache.org/namspace_mapping.html Francis On 24/05/2018 5:23 AM, Stepan Migunov wrote: > Thanks you for response, Josh! > > I got something like "Inconsistent namespace mapping properties" and > thought it was because it's impossible to set > "isNamespaceMappingEnabled" for the ODBC driver (client side). After > your explanation I understood that the "client" in this case is > queryserver but not ODBC driver. And now I need to check why queryserver > doesn't apply this property. > > -Original Message- > From: Josh Elser [mailto:els...@apache.org] > Sent: Wednesday, May 23, 2018 6:52 PM > To: user@phoenix.apache.org > Subject: Re: Phoenix ODBC driver limitations > > I'd be surprised to hear that the ODBC driver would need to know > anything about namespace-mapping. > > Do you have an error? Steps to reproduce an issue which you see? > > The reason I am surprised is that namespace mapping is an > implementation detail of the JDBC driver which lives inside of PQS -- > *not* the ODBC driver. The trivial thing you can check would be to > validate that the hbase-site.xml which PQS references is up to date > and that PQS was restarted to pick up the newest version of > hbase-site.xml > > On 5/22/18 4:16 AM, Stepan Migunov wrote: >> Hi, >> >> Is the ODBC driver from Hortonworks the only way to access Phoenix >> from .NET code now? >> The problem is that driver has some critical limitations - it seems, >> driver doesn't support Namespace Mapping (it couldn't be able to >> connect to Phoenix if phoenix.schema.isNamespaceMappingEnabled=true) >> and doesn't support query hints. >> >> Regards, >> Stepan. >>
RE: Phoenix ODBC driver limitations
Thanks you for response, Josh! I got something like "Inconsistent namespace mapping properties" and thought it was because it's impossible to set "isNamespaceMappingEnabled" for the ODBC driver (client side). After your explanation I understood that the "client" in this case is queryserver but not ODBC driver. And now I need to check why queryserver doesn't apply this property. -Original Message- From: Josh Elser [mailto:els...@apache.org] Sent: Wednesday, May 23, 2018 6:52 PM To: user@phoenix.apache.org Subject: Re: Phoenix ODBC driver limitations I'd be surprised to hear that the ODBC driver would need to know anything about namespace-mapping. Do you have an error? Steps to reproduce an issue which you see? The reason I am surprised is that namespace mapping is an implementation detail of the JDBC driver which lives inside of PQS -- *not* the ODBC driver. The trivial thing you can check would be to validate that the hbase-site.xml which PQS references is up to date and that PQS was restarted to pick up the newest version of hbase-site.xml On 5/22/18 4:16 AM, Stepan Migunov wrote: > Hi, > > Is the ODBC driver from Hortonworks the only way to access Phoenix from > .NET code now? > The problem is that driver has some critical limitations - it seems, > driver doesn't support Namespace Mapping (it couldn't be able to connect > to Phoenix if phoenix.schema.isNamespaceMappingEnabled=true) and doesn't > support query hints. > > Regards, > Stepan. >
Phoenix ODBC driver limitations
Hi, Is the ODBC driver from Hortonworks the only way to access Phoenix from .NET code now? The problem is that driver has some critical limitations - it seems, driver doesn't support Namespace Mapping (it couldn't be able to connect to Phoenix if phoenix.schema.isNamespaceMappingEnabled=true) and doesn't support query hints. Regards, Stepan.
RE: UPSERT null vlaues
Thank you James, it was “immutable”. I didn't know that it affects. *From:* James Taylor [mailto:jamestay...@apache.org] *Sent:* Friday, April 27, 2018 5:37 PM *To:* user@phoenix.apache.org *Subject:* Re: UPSERT null vlaues Hi Stepan, Please post your complete DDL and indicate the version of Phoenix and HBase you’re using. Your example should work as expected barring declaration of the table as immutable or COL2 being part of the primary key. Thanks, James On Fri, Apr 27, 2018 at 6:13 AM Stepan Migunov < stepan.migu...@firstlinesoftware.com> wrote: Hi, Could you please clarify, how I can set a value to NULL? After upsert into temp.table (ROWKEY, COL1, COL2) values (100, "ABC", null); the value of COL2 still has a previous value (COL1 has "ABC" as expected). Or there is only one way - to set STORE_NULLS = true? Thanks, Stepan.
UPSERT null vlaues
Hi, Could you please clarify, how I can set a value to NULL? After upsert into temp.table (ROWKEY, COL1, COL2) values (100, "ABC", null); the value of COL2 still has a previous value (COL1 has "ABC" as expected). Or there is only one way - to set STORE_NULLS = true? Thanks, Stepan.
RE: Storage Handler for Apache Hive
Thank you, Artem! Does it mean that Phoenix-Hive integration works with Hadoop >= 2.7 only? *From:* Artem Ervits [mailto:artemerv...@gmail.com] *Sent:* Tuesday, March 27, 2018 12:10 PM *To:* user@phoenix.apache.org *Subject:* Re: Storage Handler for Apache Hive Stepan, you're using version of Hadoop where StopWatch class is not defined https://github.com/apache/hadoop/tree/release-2.6.4-RC0/hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/util/StopWatch.java If you at least go to Hadoop 2.7, this error will disappear https://github.com/apache/hadoop/blob/release-2.7.1/hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/util/StopWatch.java On Tue, Mar 27, 2018, 4:47 AM Stepan Migunov < stepan.migu...@firstlinesoftware.com> wrote: Hi, Phoenix 4.12.0-HBase-1.1, hadoop 2.6.4, hive 2.1.1 I have setup Hive for using external Phoenix tables. But after phoenix-hive.jar was included into the hive-site.xml, the hive console give exception on some operations (e.g. show databases or query with order by clause): Exception in thread "main" java.lang.NoClassDefFoundError: org/apache/hadoop/util/StopWatch at org.apache.hadoop.mapred.FileInputFormat.getSplits(FileInputFormat.java:314) at org.apache.hadoop.hive.ql.exec.FetchOperator.getNextSplits(FetchOperator.java:372) at org.apache.hadoop.hive.ql.exec.FetchOperator.getRecordReader(FetchOperator.java:304) at org.apache.hadoop.hive.ql.exec.FetchOperator.getNextRow(FetchOperator.java:459) at org.apache.hadoop.hive.ql.exec.FetchOperator.pushRow(FetchOperator.java:428) at org.apache.hadoop.hive.ql.exec.FetchTask.fetch(FetchTask.java:146) at org.apache.hadoop.hive.ql.Driver.getResults(Driver.java:2098) at org.apache.hadoop.hive.cli.CliDriver.processLocalCmd(CliDriver.java:252) at org.apache.hadoop.hive.cli.CliDriver.processCmd(CliDriver.java:183) at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:399) at org.apache.hadoop.hive.cli.CliDriver.executeDriver(CliDriver.java:776) at org.apache.hadoop.hive.cli.CliDriver.run(CliDriver.java:714) at org.apache.hadoop.hive.cli.CliDriver.main(CliDriver.java:641) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:498) at org.apache.hadoop.util.RunJar.run(RunJar.java:221) at org.apache.hadoop.util.RunJar.main(RunJar.java:136) Caused by: java.lang.ClassNotFoundException: org.apache.hadoop.util.StopWatch at java.net.URLClassLoader.findClass(URLClassLoader.java:381) at java.lang.ClassLoader.loadClass(ClassLoader.java:424) at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:331) at java.lang.ClassLoader.loadClass(ClassLoader.java:357) ... 19 more Any suggestions are welcome.
RE: Phoenix as a source for Spark processing
The table is about 300GB in hbase. I've done some more research and now my test is very simple - I'm tryng to calculate count of records of the table. No "distincts" and etc., just phoenixTableAsDataFrame(...).count(). And now I see the issue - Spark creates about 400 task (14 executors), starts calculation, speed is pretty good. Hbase shows about 1000 requests per second. But then Sparks stops tasks as completed. I can see that Spark have read only 20% of records, but completed 50% tasks. HBase shows only 100 requests per second. When Sparks "thinks" that 99% completed (only 5 tasks left), actually it read only 70% records. The rest of work will be done by 5 tasks with 1-2 request per second... Is the any way to force Spark distribute workload evenly? I have tried to pre-split my Phonix table (now it has about 1200 regions), but it did't help. -Original Message- From: Josh Elser [mailto:els...@apache.org] Sent: Friday, March 9, 2018 2:17 AM To: user@phoenix.apache.org Subject: Re: Phoenix as a source for Spark processing How large is each row in this case? Or, better yet, how large is the table in HBase? You're spreading out approximately 7 "clients" to each Regionserver fetching results (100/14). So, you should have pretty decent saturation from Spark into HBase. I'd be taking a look at the EXPLAIN plan for your SELECT DISTINCT to really understand what Phoenix is doing. For example, are you getting ample saturation of the resources that your servers have available (32core/128Gb memory is pretty good). Validating how busy Spark is actually keeping HBase, and how much time is spent transforming the data would be good. Or, another point, are you excessively scanning data in the system which you could otherwise preclude by a different rowkey structure via logic such as a skip-scan (which would be shown in the EXPLAIN plan). You may actually find that using the built-in UPSERT SELECT logic may out-perform the Spark integration since you aren't actually doing any transformation logic inside of Spark. On 3/5/18 3:14 PM, Stepan Migunov wrote: > Hi Josh, thank you for response! > > Our cluster has 14 nodes (32 cores each/128 GB memory). The source > Phoenix table contains about 1 billion records (100 columns). We start > a Spark's job with about 100 executors. Spark executes SELECT from the > source table (select 6 columns with DISTINCT) and writes down output > to another Phoenix table. Expected that the target table will contains > about 100 million records. > HBase has 14 region servers, both tables salted with SALT_BUCKETS=42. > Spark's job running via Yarn. > > > -Original Message- > From: Josh Elser [mailto:els...@apache.org] > Sent: Monday, March 5, 2018 9:14 PM > To: user@phoenix.apache.org > Subject: Re: Phoenix as a source for Spark processing > > Hi Stepan, > > Can you better ballpark the Phoenix-Spark performance you've seen (e.g. > how much hardware do you have, how many spark executors did you use, > how many region servers)? Also, what versions of software are you using? > > I don't think there are any firm guidelines on how you can solve this > problem, but you've found the tools available for you. > > * You can try Phoenix+Spark to run over the Phoenix tables in place > * You can use Phoenix+Hive to offload the data into Hive for queries > > If Phoenix-Spark wasn't fast enough, I'd imagine using the > Phoenix-Hive integration to query the data would be similarly not fast > enough. > > It's possible that the bottleneck is something we could fix in the > integration, or fix configuration of Spark and/or Phoenix. We'd need > you to help quantify this better :) > > On 3/4/18 6:08 AM, Stepan Migunov wrote: >> In our software we need to combine fast interactive access to the >> data with quite complex data processing. I know that Phoenix intended >> for fast access, but hoped that also I could be able to use Phoenix >> as a source for complex processing with the Spark. Unfortunately, >> Phoenix + Spark shows very poor performance. E.g., querying big >> (about billion records) table with distinct takes about 2 hours. At >> the same time this task with Hive source takes a few minutes. Is it >> expected? Does it mean that Phoenix is absolutely not suitable for >> batch processing with spark and I should duplicate data to Hive and >> process it with Hive? >>
Re: Phoenix as a source for Spark processing
Some more details... We have done some simple tests to compare read/write possibility spark+hive and spark+phoenix. And now we have the following results: Copy table (with no any transformations) (about 800 million rec): Hive (TEZ) - 752 sec Spark: >From Hive to Hive: 2463 sec >From Phoenix to Hive - 13310 sec >From Hive to Phoenix - > 30240 sec We use Spark 2.2.1; hbase 1.1.2, Phonix 4.13, Hive 2.1.1 So it seems that Spark + Phoenix led great performance degradation. Any thoughts? On 2018/03/04 11:08:56, Stepan Migunov <stepan.migu...@firstlinesoftware.com> wrote: > In our software we need to combine fast interactive access to the data with > quite complex data processing. I know that Phoenix intended for fast access, > but hoped that also I could be able to use Phoenix as a source for complex > processing with the Spark. Unfortunately, Phoenix + Spark shows very poor > performance. E.g., querying big (about billion records) table with distinct > takes about 2 hours. At the same time this task with Hive source takes a few > minutes. Is it expected? Does it mean that Phoenix is absolutely not suitable > for batch processing with spark and I should duplicate data to Hive and > process it with Hive? >
RE: Phoenix as a source for Spark processing
Hi Josh, thank you for response! Our cluster has 14 nodes (32 cores each/128 GB memory). The source Phoenix table contains about 1 billion records (100 columns). We start a Spark's job with about 100 executors. Spark executes SELECT from the source table (select 6 columns with DISTINCT) and writes down output to another Phoenix table. Expected that the target table will contains about 100 million records. HBase has 14 region servers, both tables salted with SALT_BUCKETS=42. Spark's job running via Yarn. -Original Message- From: Josh Elser [mailto:els...@apache.org] Sent: Monday, March 5, 2018 9:14 PM To: user@phoenix.apache.org Subject: Re: Phoenix as a source for Spark processing Hi Stepan, Can you better ballpark the Phoenix-Spark performance you've seen (e.g. how much hardware do you have, how many spark executors did you use, how many region servers)? Also, what versions of software are you using? I don't think there are any firm guidelines on how you can solve this problem, but you've found the tools available for you. * You can try Phoenix+Spark to run over the Phoenix tables in place * You can use Phoenix+Hive to offload the data into Hive for queries If Phoenix-Spark wasn't fast enough, I'd imagine using the Phoenix-Hive integration to query the data would be similarly not fast enough. It's possible that the bottleneck is something we could fix in the integration, or fix configuration of Spark and/or Phoenix. We'd need you to help quantify this better :) On 3/4/18 6:08 AM, Stepan Migunov wrote: > In our software we need to combine fast interactive access to the data > with quite complex data processing. I know that Phoenix intended for fast > access, but hoped that also I could be able to use Phoenix as a source for > complex processing with the Spark. Unfortunately, Phoenix + Spark shows > very poor performance. E.g., querying big (about billion records) table > with distinct takes about 2 hours. At the same time this task with Hive > source takes a few minutes. Is it expected? Does it mean that Phoenix is > absolutely not suitable for batch processing with spark and I should > duplicate data to Hive and process it with Hive? >
Phoenix as a source for Spark processing
In our software we need to combine fast interactive access to the data with quite complex data processing. I know that Phoenix intended for fast access, but hoped that also I could be able to use Phoenix as a source for complex processing with the Spark. Unfortunately, Phoenix + Spark shows very poor performance. E.g., querying big (about billion records) table with distinct takes about 2 hours. At the same time this task with Hive source takes a few minutes. Is it expected? Does it mean that Phoenix is absolutely not suitable for batch processing with spark and I should duplicate data to Hive and process it with Hive?
Pool size / queue size with thin client
Hi, Could you please suggest how I can change pool size / queue size when using thin client? I have added to hbase-site.xml the following options: phoenix.query.threadPoolSize 2000 phoenix.query.queueSize 10 restarted hbase (master and regions), but still receive the following response (via JDBC-thin client): Remote driver error: RuntimeException: org.apache.phoenix.exception.PhoenixIOException: Task org.apache.phoenix.job.JobManager$InstrumentedJobFutureTask@69529e2 rejected from org.apache.phoenix.job.JobManager$1@48b8311c[Running, pool size = 128, active threads = 128, queued tasks = 5000, completed tasks = 0] My guess that settings are not applied and default values (128/5000) still used. What's wrong? Thanks, Stepan.
Spark & UpgradeInProgressException: Cluster is being concurrently upgraded from 4.11.x to 4.12.x
Hi, I have just upgraded my cluster to Phoenix 4.12 and got an issue with tasks running on Spark 2.2 (yarn cluster mode). Any attempts to use method phoenixTableAsDataFrame to load data from existing database causes an exception (see below). The tasks worked fine on version 4.11. I have checked connection with sqlline - it works now and shows that version is 4.12. Moreover, I have noticed, that if limit the number of executors to one, the Spark's task executing successfully too! It looks like that executors running in parallel "interferes" each other’s and could not acquire version's mutex. Any suggestions please? *final Connection connection = ConnectionUtil.getInputConnection(configuration, overridingProps);* *User class threw exception: org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 0.0 failed 4 times, most recent failure: Lost task 0.3 in stage 0.0 (TID 36, n7701-hdp005, executor 26): java.lang.RuntimeException: org.apache.phoenix.exception.UpgradeInProgressException: Cluster is being concurrently upgraded from 4.11.x to 4.12.x. Please retry establishing connection.* *at org.apache.phoenix.mapreduce.PhoenixInputFormat.getQueryPlan(PhoenixInputFormat.java:201)* *at org.apache.phoenix.mapreduce.PhoenixInputFormat.createRecordReader(PhoenixInputFormat.java:76)* *at org.apache.spark.rdd.NewHadoopRDD$$anon$1.liftedTree1$1(NewHadoopRDD.scala:180)* *at org.apache.spark.rdd.NewHadoopRDD$$anon$1.(NewHadoopRDD.scala:179)* *at org.apache.spark.rdd.NewHadoopRDD.compute(NewHadoopRDD.scala:134)* *at org.apache.spark.rdd.NewHadoopRDD.compute(NewHadoopRDD.scala:69)* *at org.apache.phoenix.spark.PhoenixRDD.compute(PhoenixRDD.scala:64)* *at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)* *at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)* *at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)* *at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)* *at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)* *at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)* *at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)* *at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)* *at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)* *at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)* *at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)* *at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)* *at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)* *at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)* *at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)* *at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)* *at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)* *at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:96)* *at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:53)* *at org.apache.spark.scheduler.Task.run(Task.scala:108)* *at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:335)* *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)* *Caused by: org.apache.phoenix.exception.UpgradeInProgressException: Cluster is being concurrently upgraded from 4.11.x to 4.12.x. Please retry establishing connection.* *at org.apache.phoenix.query.ConnectionQueryServicesImpl.acquireUpgradeMutex(ConnectionQueryServicesImpl.java:3173)* *at org.apache.phoenix.query.ConnectionQueryServicesImpl.upgradeSystemTables(ConnectionQueryServicesImpl.java:2567)* *at org.apache.phoenix.query.ConnectionQueryServicesImpl$12.call(ConnectionQueryServicesImpl.java:2440)* *at org.apache.phoenix.query.ConnectionQueryServicesImpl$12.call(ConnectionQueryServicesImpl.java:2360)* *at org.apache.phoenix.util.PhoenixContextExecutor.call(PhoenixContextExecutor.java:76)* *at org.apache.phoenix.query.ConnectionQueryServicesImpl.init(ConnectionQueryServicesImpl.java:2360)* *at org.apache.phoenix.jdbc.PhoenixDriver.getConnectionQueryServices(PhoenixDriver.java:255)* *at org.apache.phoenix.jdbc.PhoenixEmbeddedDriver.createConnection(PhoenixEmbeddedDriver.java:150)* *at org.apache.phoenix.jdbc.PhoenixDriver.connect(PhoenixDriver.java:221)* *at java.sql.DriverManager.getConnection(DriverManager.java:664)* *at java.sql.DriverManager.getConnection(DriverManager.java:208)* *at org.apache.phoenix.mapreduce.util.ConnectionUtil.getConnection(ConnectionUtil.java:98)* *at org.apache.phoenix.mapreduce.util.ConnectionUtil.getInputConnection(ConnectionUtil.java:57)* *at org.apache.phoenix.mapreduce.PhoenixInputFormat.getQueryPlan(PhoenixInputFormat.java:176)* *... 30 more*