[jira] [Commented] (SPARK-19532) [Core]`DataStreamer for file` threads of DFSOutputStream leak if set `spark.speculation` to true
[ https://issues.apache.org/jira/browse/SPARK-19532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15999016#comment-15999016 ] Abhishek Madav commented on SPARK-19532: I am running into this issue wherein codepath similar to hiveWriterContainer is trying to the HDFS location. I tried setting spark.speculation to false but it doesn't seem to be the issue. Is there any workaround? This wait-time leads to make the job run real slow. > [Core]`DataStreamer for file` threads of DFSOutputStream leak if set > `spark.speculation` to true > > > Key: SPARK-19532 > URL: https://issues.apache.org/jira/browse/SPARK-19532 > Project: Spark > Issue Type: Bug > Components: Spark Core, SQL >Affects Versions: 2.1.0 >Reporter: StanZhai >Priority: Critical > > When set `spark.speculation` to true, from thread dump page of Executor of > WebUI, I found that there are about 1300 threads named "DataStreamer for > file > /test/data/test_temp/_temporary/0/_temporary/attempt_20170207172435_80750_m_69_1/part-00069-690407af-0900-46b1-9590-a6d6c696fe68.snappy.parquet" > in TIMED_WAITING state. > {code} > java.lang.Object.wait(Native Method) > org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:564) > {code} > The off-heap memory exceeds a lot until Executor exited with OOM exception. > This problem occurs only when writing data to the Hadoop(tasks may be killed > by Executor during writing). > Could this be related to [https://issues.apache.org/jira/browse/HDFS-9812]? > The version of Hadoop is 2.6.4. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-19532) [Core]`DataStreamer for file` threads of DFSOutputStream leak if set `spark.speculation` to true
[ https://issues.apache.org/jira/browse/SPARK-19532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15952730#comment-15952730 ] Hyukjin Kwon commented on SPARK-19532: -- It sounds no one except for the reporter could reproduce this. If you are going to fix it, that's great but if not, I think it might be better to resolve this for now. Anyone can reopen this with more details and symptoms if anyone meets this issue. I guess this will probably be open without more actions. > [Core]`DataStreamer for file` threads of DFSOutputStream leak if set > `spark.speculation` to true > > > Key: SPARK-19532 > URL: https://issues.apache.org/jira/browse/SPARK-19532 > Project: Spark > Issue Type: Bug > Components: Spark Core, SQL >Affects Versions: 2.1.0 >Reporter: StanZhai >Priority: Critical > > When set `spark.speculation` to true, from thread dump page of Executor of > WebUI, I found that there are about 1300 threads named "DataStreamer for > file > /test/data/test_temp/_temporary/0/_temporary/attempt_20170207172435_80750_m_69_1/part-00069-690407af-0900-46b1-9590-a6d6c696fe68.snappy.parquet" > in TIMED_WAITING state. > {code} > java.lang.Object.wait(Native Method) > org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:564) > {code} > The off-heap memory exceeds a lot until Executor exited with OOM exception. > This problem occurs only when writing data to the Hadoop(tasks may be killed > by Executor during writing). > Could this be related to [https://issues.apache.org/jira/browse/HDFS-9812]? > The version of Hadoop is 2.6.4. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-19532) [Core]`DataStreamer for file` threads of DFSOutputStream leak if set `spark.speculation` to true
[ https://issues.apache.org/jira/browse/SPARK-19532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15864953#comment-15864953 ] StanZhai commented on SPARK-19532: -- I can reproduce this by split our online data to the production test cluster use our Spark application. Our application is a web service, sql job requests are concurrently handled by it(like hive-thriftserver). It's really a bit difficult to reproduce in the development environment. > [Core]`DataStreamer for file` threads of DFSOutputStream leak if set > `spark.speculation` to true > > > Key: SPARK-19532 > URL: https://issues.apache.org/jira/browse/SPARK-19532 > Project: Spark > Issue Type: Bug > Components: Spark Core, SQL >Affects Versions: 2.1.0 >Reporter: StanZhai >Priority: Critical > > When set `spark.speculation` to true, from thread dump page of Executor of > WebUI, I found that there are about 1300 threads named "DataStreamer for > file > /test/data/test_temp/_temporary/0/_temporary/attempt_20170207172435_80750_m_69_1/part-00069-690407af-0900-46b1-9590-a6d6c696fe68.snappy.parquet" > in TIMED_WAITING state. > {code} > java.lang.Object.wait(Native Method) > org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:564) > {code} > The off-heap memory exceeds a lot until Executor exited with OOM exception. > This problem occurs only when writing data to the Hadoop(tasks may be killed > by Executor during writing). > Could this be related to [https://issues.apache.org/jira/browse/HDFS-9812]? > The version of Hadoop is 2.6.4. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-19532) [Core]`DataStreamer for file` threads of DFSOutputStream leak if set `spark.speculation` to true
[ https://issues.apache.org/jira/browse/SPARK-19532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15863340#comment-15863340 ] Herman van Hovell commented on SPARK-19532: --- I understand that this is a blocker for you; however this is not a release blocker. The problem with this ticket is that it is almost impossible to reproduce. Is there any chance that you create a reproducible error? > [Core]`DataStreamer for file` threads of DFSOutputStream leak if set > `spark.speculation` to true > > > Key: SPARK-19532 > URL: https://issues.apache.org/jira/browse/SPARK-19532 > Project: Spark > Issue Type: Bug > Components: Spark Core, SQL >Affects Versions: 2.1.0 >Reporter: StanZhai >Priority: Critical > > When set `spark.speculation` to true, from thread dump page of Executor of > WebUI, I found that there are about 1300 threads named "DataStreamer for > file > /test/data/test_temp/_temporary/0/_temporary/attempt_20170207172435_80750_m_69_1/part-00069-690407af-0900-46b1-9590-a6d6c696fe68.snappy.parquet" > in TIMED_WAITING state. > {code} > java.lang.Object.wait(Native Method) > org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:564) > {code} > The off-heap memory exceeds a lot until Executor exited with OOM exception. > This problem occurs only when writing data to the Hadoop(tasks may be killed > by Executor during writing). > Could this be related to [https://issues.apache.org/jira/browse/HDFS-9812]? > The version of Hadoop is 2.6.4. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-19532) [Core]`DataStreamer for file` threads of DFSOutputStream leak if set `spark.speculation` to true
[ https://issues.apache.org/jira/browse/SPARK-19532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15862625#comment-15862625 ] StanZhai commented on SPARK-19532: -- We have been trying to upgrade our Spark from the releasing of Spark 2.1.0. This version is not available for us because of the memory problems. > [Core]`DataStreamer for file` threads of DFSOutputStream leak if set > `spark.speculation` to true > > > Key: SPARK-19532 > URL: https://issues.apache.org/jira/browse/SPARK-19532 > Project: Spark > Issue Type: Bug > Components: Spark Core, SQL >Affects Versions: 2.1.0 >Reporter: StanZhai >Priority: Blocker > > When set `spark.speculation` to true, from thread dump page of Executor of > WebUI, I found that there are about 1300 threads named "DataStreamer for > file > /test/data/test_temp/_temporary/0/_temporary/attempt_20170207172435_80750_m_69_1/part-00069-690407af-0900-46b1-9590-a6d6c696fe68.snappy.parquet" > in TIMED_WAITING state. > {code} > java.lang.Object.wait(Native Method) > org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:564) > {code} > The off-heap memory exceeds a lot until Executor exited with OOM exception. > This problem occurs only when writing data to the Hadoop(tasks may be killed > by Executor during writing). -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-19532) [Core]`DataStreamer for file` threads of DFSOutputStream leak if set `spark.speculation` to true
[ https://issues.apache.org/jira/browse/SPARK-19532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15862443#comment-15862443 ] Hyukjin Kwon commented on SPARK-19532: -- Is it a blocker BTW? > [Core]`DataStreamer for file` threads of DFSOutputStream leak if set > `spark.speculation` to true > > > Key: SPARK-19532 > URL: https://issues.apache.org/jira/browse/SPARK-19532 > Project: Spark > Issue Type: Bug > Components: Spark Core, SQL >Affects Versions: 2.1.0 >Reporter: StanZhai >Priority: Blocker > > When set `spark.speculation` to true, from thread dump page of Executor of > WebUI, I found that there are about 1300 threads named "DataStreamer for > file > /test/data/test_temp/_temporary/0/_temporary/attempt_20170207172435_80750_m_69_1/part-00069-690407af-0900-46b1-9590-a6d6c696fe68.snappy.parquet" > in TIMED_WAITING state. > {code} > java.lang.Object.wait(Native Method) > org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:564) > {code} > The off-heap memory exceeds a lot until Executor exited with OOM exception. > This problem occurs only when writing data to the Hadoop(tasks may be killed > by Executor during writing). -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org