About darouwan

darouwan · ‎04-26-2019

I have met the similar problem. I have move the block pool to another place, but the problem occurs as well. How to fix it? su -l hdfs -c "/usr/hdp/current/hadoop-hdfs-datanode/../hadoop/sbin/hadoop-daemon.sh start datanode" output following message: Java Hotshpot 64bit server VM warning: cannot open file /var/log/hadoop/hdfs/gc.log-2019... due to No such file or directory

darouwan · ‎03-12-2019

Thanks @Jordan Moore As 'Kafka is very memory and disk sensitive. ', do you recommend to install kafka brokers on a virtual machine, as I cannot have more dedicated machines for kafka?

darouwan · ‎03-11-2019

I meet a significant performance problem recently. I have about 30 spark streaming applications, which read data from kafka and write the data to hdfs. But recently the writing progress on some spark executor become very slow. The data amount for each spark tasks are similar, but the time cost of tasks are in great difference, where the slowest one is about 4 times of the fastest one. I have checked the disk usage, where disk use time on some hosts are about 80% to 90%. So I guess if it is caused by slow hdfs writing speed, because of my kafka broker, hdfs data node, yarn nodemanager locating on same hosts. So will it actually affect the performance?

darouwan · ‎12-07-2018

Hi @Jagadeesan A S my current save mode is append. My sparking streaming apps will run every 5 min, it is not convenient to delete manually....So I think the better solution is customize the temp location. Or Can I set offset of the scheduled running time? For example, my current 2 apps every 5 minutes, that's run at 0, 5, 10, 15, 20 Can I set a schedule, make one still runs at 0, 5, 10 , 15, and another runs at 2.5, 7.5, 10.5?

darouwan · ‎12-06-2018

Thanks @Jagadeesan A S _temporary is a temp directory under path of the df.write.parquet(path) on hdfs. However spark.local.dir default value is /tmp, and in document, Directory to use for "scratch" space in Spark, including map output files and RDDs that get stored on disk. This should be on a fast, local disk in your system. So it should be a directory on local file system. I am not sure spark.local.dir refers to the temp directory of spark writing ...

darouwan · ‎12-06-2018

I have two spark applications writing data to one directory on HDFS, which cause the faster completed app will delete the working directory _temporary containing some temp file belonging to another app. So can I specify a _temporary directory for each Spark application？

darouwan · ‎11-30-2018

I use kafka with default value(PLAINTEXT://localhost:9092), however, it broadcasts it's hostname instead of ip address. The result is, if the producer runs a host without the kafka host info, it cannot send messages to kafka. How to resolve it?

darouwan · ‎11-20-2018

OK @Jay Kumar SenSharma So can I upgrade the ambari metrics service individually , besides to the whole Ambari?

darouwan · ‎11-20-2018

for netstat -anlp | grep :6188 | grep CLOSE_WAIT | wc -l I get 0. May my restarting the service solve this problem temporarily. You mean the CLOSE_WAITING problem relates to Failed to invoke 'unbuffer' method inclassclass org.apache.hadoop.fs.FSDataInputStream ?

darouwan · ‎11-20-2018

I am using Ambari 2.6.1.0 The ambari metrics collector log is stored on production environment. I cannot export it. And I can give you the error log by hand typing copy: now In embedded mode, ambari metrics log: MetaDataProtos$MetaDataService for row \x00\x00METRIC_RECORD .... Caused by java.lang.InterruptedException ams hbase log: FSDataInputStreamWrapper: Failed to invoke 'unbuffer' method in class class org.apache.hadoop.fs.FSDataInputStream So there may be a TCP socket connection left open in CLOSE_WAIT state .... caused by java.lang.UnsupportedOperationException: this stream does not support ubbuffering All settings for ams are in default. There are 6 nodes in total in cluster. The host running ams has 64 cores and 256GB. Currently it has 53GB free memory and 233GB free memory in cache

darouwan · ‎11-19-2018

my ambari metric collector always doesn't work. In log, I found Failed to get result with timeout, timeout = 300000ms row 'METRIC_AGGREGATE' on table 'SYSTEM.CATALOG' at region=SYSTEM.CATALOG, ***** host name=[hostname], [port] *** seNum=6....causes by ....IOException: Failed to get result with timeout, timeout=300000ms then stuck at org.apache.hadoop.hbase.client.AsynProcess:#1 waiting for 28379 actions to finish It seems relate to hbase, but my hbase runs well without any error If I restart ambari metric collector, it recovers immediately, and become unavailable again after several hours. How to fix it? Thanks

darouwan · ‎11-13-2018

Thanks @KB And another question: When my spark application writing massive of data to hdfs, it always throws error message like following: No lease on /user/xx/sample_2016/_temporary/0/_temporary/attempt_201604141035_0058_m_019029_0/part-r-19029-1b93e1fa-9284-4f2c-821a-c83795ad27c1.gz.parquet:File does not exist.HolderDFSClient_NONMAPREDUCE_1239207978_115 does not have any open files. How to solve this problem? I search online and others said it is related to dfs.datanode.max.xcievers

darouwan · ‎11-13-2018

I increase the ipc max length according to this https://community.hortonworks.com/questions/101841/issue-requested-data-length-146629817-is-longer-th.html The hdfs service seems back to work.

darouwan · ‎11-13-2018

Thanks @KB I have reset the dfs.datanode.max.xcievers to 32768, is it still too high? I increase it to avoid "No lease on file (inode 5425306)" error. So what's the proper value for this property? If I set the value to a proper value, will the missing block be recovered automatically?

darouwan · ‎11-13-2018

Yesterday I add three more data nodes to my hdfs cluster with hdp 2.6.4. Few hours later, because of sparking writing error(No lease on...), I increase dfs.datanode.max.xcievers to 65536 and increase the heap size of name node and data node from 5G to 12G. And then restart it. However, the hdfs restart progress pauses in name node stage. It shows it is always in safe mode, and continues for 10 minutes. I force to leave the safe mode manually, and then hdfs reports a lot blocks are missing(about more than 90%). I checked the log of datanode and namenode, there are two kinds of error log: 1. In name node: Requested data length ** is longer than maximum configured RPC length ** 2. In data node: End of file exception between local host is "***", destination host is "**:8020" So how can I recovery my missing file? and what's the actual cause of this problem?

darouwan · ‎10-15-2018

Solve by using HttpFs. It set a gateway where no need to access data node.

darouwan · ‎10-11-2018

I need to read files on hdfs which is distributed on another network, where I can access it via proxy. Now I can access the directories, files status via WebHDFS by specific port successfully. However when I try to read the file content, it redirects me to another data node url, while I only have access to name node. So how can I read the file in a convenient way?

darouwan · ‎10-10-2018

@Aditya Sirna So in default , there are up to 1000 lines of results stored on hdfs for each query? If I increase the limit, will it have some negative effects? Such as slow http transferring? Or result receiving failed?

darouwan · ‎10-10-2018

@Aditya Sirna Thanks Aditya So what about paging? Since the whole results are saved on hdfs in JSON format, if I need to load part of whole result, just load the whole json file and cut out part of it by given page size and page number in memory ? In practice for zeppelin, will it have out of memory problem if the size is too huge?

darouwan · ‎10-10-2018

I am working on designing a hdfs query system based on spark, which containing a paging function, and zeppelin seems be a good sample for me. Now I have a problem. I see spark or spark sql query results are existed even I refresh or reopen the notebook. So the results must be saved on some place. So I am wondering where these result data is saved on? If the data is saved on database, what if the result data size is pretty huge so that causing the database performance problem?

darouwan · ‎08-16-2018

Hi @Jonathan Sneep Fine. thanks. I have added the user and group info to my namenode. So the typical way to adding the new user or group is creating the user and group on namenode, and waiting for usersync to sync the user info to Ranger? So if I don't care the group policy, creating internal user in ranger and specifying them in allow conditions also works? At least it seems work in practice..

darouwan · ‎08-16-2018

Hi @Jonathan Sneep Not yet. So I need to add the user the related group on my namenode host manually?

darouwan · ‎08-15-2018

Hi, @Jonathan Sneep Thanks for your response. Actually both user and group are created in Ranger, which are internal for ranger

darouwan · ‎08-15-2018

I meet some problem in ranger authentication. Here is my step to represent it: 1. I create one account in ranger, where the username is test01 2. I set it belong to a group test_group01 3. In the ranger hdfs policy, I set test_group01 has access to the directory /data/ If it runs normally, the test01 user should have the access to /data/ from the privilege inheriting from the group "test_group01". But in practice, it cannot access the directory /data. However if I specify the 'select user' with the test01, it works well. So it seems that specifying the group in policy doesn't work, and specifying the permitted user is fine. How to solve it? Thanks!

darouwan · ‎07-12-2018

Thanks Jay. I checked curl and libcurl version by running "yum list | grep curl", their version is . curl.x86_64 7.19.7-46.el6 libcurl.x86_64 7.19.7-46.el6 python-pycurl.x86_64 7.19.0-8.el6 libcurl.i686 7.19.7-46.el6 libcurl-devel.i686 7.19.7-46.el6 libcurl-devel.x86_64 7.19.7-46.el6 curl -V prints the following info: curl 7.19.7(x86-64-redhat-linux-gnu) libcurl/7.19.7... Protocals:... Features: GSS-Negotiate ... If I run the alert_spark2_livy_port.py script independently, it runs well What confuse me is, all my three hosts have the complete same version of curl, but only one have the above problem.

darouwan · ‎07-12-2018

The spark livy alert always reports an alert : Connection failed on host ***:8999 In detail, it prints ExecutionFailed: Execution of 'curl -s o /dev/null -w'%{http_code} --negotiate -u: -k http://host:8999/session | grep 200' return 1, curl: option --negotiate: the installed liburl version doesn't support this curl: try curl --help... I have 3 host in this cluster, but only one host report this alert. I have checked the curl version and libcurl on hosts respectively, and they are all same. It may caused by installing anaconda and python version changing, but I am not sure as default python version is 2.6. How to fix it? Thanks!

darouwan · ‎04-03-2018

I am trying to read data from kafka and writing them in parquet format via Spark Streaming. The problem is, the data from kafka are in variable data structure. For example, app one has columns A,B,C, app two has columns B,C,D. So the data frame I read from kafka has all columns ABCD. When I decide to write the dataframe to parquet file partitioned with app name, the parquet file of app one also contains columns D, where the columns D is empty and it contains no data actually. So how to filter the empty columns when I writing dataframe to parquet? Thanks!

darouwan · ‎04-01-2018

I have submitted a spark java program via "Spark Submit Jar" and it looks running well. However, when I click the logs link in specified application in Job tab in hue, it shows "cannot acces: /jobbrowser/jobs/appliacation_****/single_logs." So how can I find logs of running spark application?

darouwan · ‎03-26-2018

Fixed it by recovering spark home setting

darouwan · ‎03-23-2018

I am testing spark within zeppelin. But in running tutorial %spark2.spark spark.verson It throws the following error: java.lang.NullPointerException at org.apache.zeppelin.spark.Utils.invokeMethod(Utils.java:38) at org.apache.zeppelin.spark.Utils.invokeMethod(Utils.java:33) at org.apache.zeppelin.spark.SparkInterpreter.createSparkContext_2(SparkInterpreter.java:391) at org.apache.zeppelin.spark.SparkInterpreter.createSparkContext(SparkInterpreter.java:380) at org.apache.zeppelin.spark.SparkInterpreter.getSparkContext(SparkInterpreter.java:146) at org.apache.zeppelin.spark.SparkInterpreter.open(SparkInterpreter.java:828) at org.apache.zeppelin.interpreter.LazyOpenInterpreter.open(LazyOpenInterpreter.java:70) at org.apache.zeppelin.interpreter.remote.RemoteInterpreterServer$InterpretJob.jobRun(RemoteInterpreterServer.java:483) at org.apache.zeppelin.scheduler.Job.run(Job.java:175) at org.apache.zeppelin.scheduler.FIFOScheduler$1.run(FIFOScheduler.java:139) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) 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) Then I disable the hive context according to https://stackoverflow.com/questions/43289067/getting-nullpointerexception-when-running-spark-code-in-zeppelin-0-7-1 , the same exception is still thrown. How to solve it? ========================================================= Update 1: I have checked the spark interpreter log, and get the following error: requirement failed:/python/lib/pyspark.zip not found;cannot run pyspark application in YARN mode. How to locate this file or config the path?

Online	Offline
Last Visited	‎04-02-2018 12:02 AM

Date Registered	‎12-21-2017 12:43 AM
Last Visited	‎04-02-2018 12:02 AM
Total Messages Posted	67
Total Kudos Received	3

Re: Access file on hdfs via proxy

Re: NullPointerException when running spark on Zep...

Re: datanode not starting

Re: Can kafka broker, yarn nodemanager, hdfs data ...

Can kafka broker, yarn nodemanager, hdfs data node...

Re: How to change Spark _temporary directory when ...

Re: How to change Spark _temporary directory when ...

How to change Spark _temporary directory when writ...

How to set kafka to broadcast the ip address rathe...

Re: No data available for Ambari Metric Collector

Re: No data available for Ambari Metric Collector

Re: No data available for Ambari Metric Collector

No data available for Ambari Metric Collector

Re: A lot of blocks missing in HDFS

Re: A lot of blocks missing in HDFS

Re: A lot of blocks missing in HDFS

A lot of blocks missing in HDFS

Re: Access file on hdfs via proxy

Access file on hdfs via proxy

Re: How does zeppelin storage query result?

Re: How does zeppelin storage query result?

How does zeppelin storage query result?

Re: Ranger auth error in hdfs

Re: Ranger auth error in hdfs

Re: Ranger auth error in hdfs

Ranger auth error in hdfs

Re: "installed libcurl version doesn't support thi...

"installed libcurl version doesn't support this cu...

How to delete empty columns in df when writing to ...

How to view the log of submitted spark jar program...

Re: NullPointerException when running spark on Zep...

NullPointerException when running spark on Zeppeli...