[jira] [Reopened] (SPARK-23130) Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout)
[ https://issues.apache.org/jira/browse/SPARK-23130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts reopened SPARK-23130: -- > Spark Thrift does not clean-up temporary files (/tmp/*_resources and > /tmp/hive/*.pipeout) > - > > Key: SPARK-23130 > URL: https://issues.apache.org/jira/browse/SPARK-23130 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 1.6.3, 2.1.0, 2.2.0 > Environment: * Hadoop distributions: HDP 2.5 - 2.6.3.0 > * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 >Reporter: Sean Roberts >Priority: Major > Labels: thrift > > Spark Thrift is not cleaning up /tmp for files & directories named like: > /tmp/hive/*.pipeout > /tmp/*_resources > There are such a large number that /tmp quickly runs out of inodes *causing > the partition to be unusable and many services to crash*. This is even true > when the only jobs submitted are routine service checks. > Used `strace` to show that Spark Thrift is responsible: > {code:java} > strace.out.118864:04:53:49 > open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", > O_RDWR|O_CREAT|O_EXCL, 0666) = 134 > strace.out.118864:04:53:49 > mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 > {code} > *Those files were left behind, even days later.* > > Example files: > {code:java} > # stat > /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout > File: > ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ > Size: 0 Blocks: 0 IO Block: 4096 regular empty file > Device: fe09h/65033d Inode: 678 Links: 1 > Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:53:49.126777260 -0600 > Modify: 2017-12-19 04:53:49.126777260 -0600 > Change: 2017-12-19 04:53:49.126777260 -0600 > Birth: - > # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources > File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ > Size: 4096 Blocks: 8 IO Block: 4096 directory > Device: fe09h/65033d Inode: 668 Links: 2 > Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:57:38.458937635 -0600 > Modify: 2017-12-19 04:53:49.062777216 -0600 > Change: 2017-12-19 04:53:49.066777218 -0600 > Birth: - > {code} > Showing the large number: > {code:java} > # find /tmp/ -name '*_resources' | wc -l > 68340 > # find /tmp/hive -name "*.pipeout" | wc -l > 51837 > {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-23130) Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout)
[ https://issues.apache.org/jira/browse/SPARK-23130?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16329464#comment-16329464 ] Sean Roberts commented on SPARK-23130: -- Marco - Which JIRA resolves the pipeout issue or incorporates the pipeout fixes into Spark Thrift? > Spark Thrift does not clean-up temporary files (/tmp/*_resources and > /tmp/hive/*.pipeout) > - > > Key: SPARK-23130 > URL: https://issues.apache.org/jira/browse/SPARK-23130 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 1.6.3, 2.1.0, 2.2.0 > Environment: * Hadoop distributions: HDP 2.5 - 2.6.3.0 > * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 >Reporter: Sean Roberts >Priority: Major > Labels: thrift > > Spark Thrift is not cleaning up /tmp for files & directories named like: > /tmp/hive/*.pipeout > /tmp/*_resources > There are such a large number that /tmp quickly runs out of inodes *causing > the partition to be unusable and many services to crash*. This is even true > when the only jobs submitted are routine service checks. > Used `strace` to show that Spark Thrift is responsible: > {code:java} > strace.out.118864:04:53:49 > open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", > O_RDWR|O_CREAT|O_EXCL, 0666) = 134 > strace.out.118864:04:53:49 > mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 > {code} > *Those files were left behind, even days later.* > > Example files: > {code:java} > # stat > /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout > File: > ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ > Size: 0 Blocks: 0 IO Block: 4096 regular empty file > Device: fe09h/65033d Inode: 678 Links: 1 > Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:53:49.126777260 -0600 > Modify: 2017-12-19 04:53:49.126777260 -0600 > Change: 2017-12-19 04:53:49.126777260 -0600 > Birth: - > # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources > File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ > Size: 4096 Blocks: 8 IO Block: 4096 directory > Device: fe09h/65033d Inode: 668 Links: 2 > Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:57:38.458937635 -0600 > Modify: 2017-12-19 04:53:49.062777216 -0600 > Change: 2017-12-19 04:53:49.066777218 -0600 > Birth: - > {code} > Showing the large number: > {code:java} > # find /tmp/ -name '*_resources' | wc -l > 68340 > # find /tmp/hive -name "*.pipeout" | wc -l > 51837 > {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-23130) Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout)
[ https://issues.apache.org/jira/browse/SPARK-23130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts updated SPARK-23130: - Labels: thrift (was: ) > Spark Thrift does not clean-up temporary files (/tmp/*_resources and > /tmp/hive/*.pipeout) > - > > Key: SPARK-23130 > URL: https://issues.apache.org/jira/browse/SPARK-23130 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 1.6.3, 2.1.0, 2.2.0 > Environment: * Hadoop distributions: HDP 2.5 - 2.6.3.0 > * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 >Reporter: Sean Roberts >Priority: Major > Labels: thrift > > Spark Thrift is not cleaning up /tmp for files & directories named like: > /tmp/hive/*.pipeout > /tmp/*_resources > There are such a large number that /tmp quickly runs out of inodes *causing > the partition to be unusable and many services to crash*. This is even true > when the only jobs submitted are routine service checks. > Used `strace` to show that Spark Thrift is responsible: > {code:java} > strace.out.118864:04:53:49 > open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", > O_RDWR|O_CREAT|O_EXCL, 0666) = 134 > strace.out.118864:04:53:49 > mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 > {code} > *Those files were left behind, even days later.* > > Example files: > {code:java} > # stat > /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout > File: > ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ > Size: 0 Blocks: 0 IO Block: 4096 regular empty file > Device: fe09h/65033d Inode: 678 Links: 1 > Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:53:49.126777260 -0600 > Modify: 2017-12-19 04:53:49.126777260 -0600 > Change: 2017-12-19 04:53:49.126777260 -0600 > Birth: - > # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources > File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ > Size: 4096 Blocks: 8 IO Block: 4096 directory > Device: fe09h/65033d Inode: 668 Links: 2 > Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:57:38.458937635 -0600 > Modify: 2017-12-19 04:53:49.062777216 -0600 > Change: 2017-12-19 04:53:49.066777218 -0600 > Birth: - > {code} > Showing the large number: > {code:java} > # find /tmp/ -name '*_resources' | wc -l > 68340 > # find /tmp/hive -name "*.pipeout" | wc -l > 51837 > {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-23130) Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout)
[ https://issues.apache.org/jira/browse/SPARK-23130?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16328568#comment-16328568 ] Sean Roberts commented on SPARK-23130: -- * SPARK-15401: Similar report for the "_resources" files * HIVE-6091: Possibly a fix for the "pipeout" files > Spark Thrift does not clean-up temporary files (/tmp/*_resources and > /tmp/hive/*.pipeout) > - > > Key: SPARK-23130 > URL: https://issues.apache.org/jira/browse/SPARK-23130 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 1.6.3, 2.1.0, 2.2.0 > Environment: * Hadoop distributions: HDP 2.5 - 2.6.3.0 > * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 >Reporter: Sean Roberts >Priority: Major > > Spark Thrift is not cleaning up /tmp for files & directories named like: > /tmp/hive/*.pipeout > /tmp/*_resources > There are such a large number that /tmp quickly runs out of inodes *causing > the partition to be unusable and many services to crash*. This is even true > when the only jobs submitted are routine service checks. > Used `strace` to show that Spark Thrift is responsible: > {code:java} > strace.out.118864:04:53:49 > open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", > O_RDWR|O_CREAT|O_EXCL, 0666) = 134 > strace.out.118864:04:53:49 > mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 > {code} > *Those files were left behind, even days later.* > ** > > Example files: > {code:java} > # stat > /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout > File: > ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ > Size: 0 Blocks: 0 IO Block: 4096 regular empty file > Device: fe09h/65033d Inode: 678 Links: 1 > Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:53:49.126777260 -0600 > Modify: 2017-12-19 04:53:49.126777260 -0600 > Change: 2017-12-19 04:53:49.126777260 -0600 > Birth: - > # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources > File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ > Size: 4096 Blocks: 8 IO Block: 4096 directory > Device: fe09h/65033d Inode: 668 Links: 2 > Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:57:38.458937635 -0600 > Modify: 2017-12-19 04:53:49.062777216 -0600 > Change: 2017-12-19 04:53:49.066777218 -0600 > Birth: - > {code} > Showing the large number: > {code:java} > # find /tmp/ -name '*_resources' | wc -l > 68340 > # find /tmp/hive -name "*.pipeout" | wc -l > 51837 > {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-23130) Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout)
[ https://issues.apache.org/jira/browse/SPARK-23130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts updated SPARK-23130: - Description: Spark Thrift is not cleaning up /tmp for files & directories named like: /tmp/hive/*.pipeout /tmp/*_resources There are such a large number that /tmp quickly runs out of inodes *causing the partition to be unusable and many services to crash*. This is even true when the only jobs submitted are routine service checks. Used `strace` to show that Spark Thrift is responsible: {code:java} strace.out.118864:04:53:49 open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", O_RDWR|O_CREAT|O_EXCL, 0666) = 134 strace.out.118864:04:53:49 mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 {code} *Those files were left behind, even days later.* Example files: {code:java} # stat /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout File: ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ Size: 0 Blocks: 0 IO Block: 4096 regular empty file Device: fe09h/65033dInode: 678 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) Access: 2017-12-19 04:53:49.126777260 -0600 Modify: 2017-12-19 04:53:49.126777260 -0600 Change: 2017-12-19 04:53:49.126777260 -0600 Birth: - # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ Size: 4096Blocks: 8 IO Block: 4096 directory Device: fe09h/65033dInode: 668 Links: 2 Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) Access: 2017-12-19 04:57:38.458937635 -0600 Modify: 2017-12-19 04:53:49.062777216 -0600 Change: 2017-12-19 04:53:49.066777218 -0600 Birth: - {code} Showing the large number: {code:java} # find /tmp/ -name '*_resources' | wc -l 68340 # find /tmp/hive -name "*.pipeout" | wc -l 51837 {code} was: Spark Thrift is not cleaning up /tmp for files & directories named like: /tmp/hive/*.pipeout /tmp/*_resources There are such a large number that /tmp quickly runs out of inodes *causing the partition to be unusable and many services to crash*. This is even true when the only jobs submitted are routine service checks. Used `strace` to show that Spark Thrift is responsible: {code:java} strace.out.118864:04:53:49 open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", O_RDWR|O_CREAT|O_EXCL, 0666) = 134 strace.out.118864:04:53:49 mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 {code} *Those files were left behind, even days later.* ** Example files: {code:java} # stat /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout File: ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ Size: 0 Blocks: 0 IO Block: 4096 regular empty file Device: fe09h/65033dInode: 678 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) Access: 2017-12-19 04:53:49.126777260 -0600 Modify: 2017-12-19 04:53:49.126777260 -0600 Change: 2017-12-19 04:53:49.126777260 -0600 Birth: - # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ Size: 4096Blocks: 8 IO Block: 4096 directory Device: fe09h/65033dInode: 668 Links: 2 Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) Access: 2017-12-19 04:57:38.458937635 -0600 Modify: 2017-12-19 04:53:49.062777216 -0600 Change: 2017-12-19 04:53:49.066777218 -0600 Birth: - {code} Showing the large number: {code:java} # find /tmp/ -name '*_resources' | wc -l 68340 # find /tmp/hive -name "*.pipeout" | wc -l 51837 {code} > Spark Thrift does not clean-up temporary files (/tmp/*_resources and > /tmp/hive/*.pipeout) > - > > Key: SPARK-23130 > URL: https://issues.apache.org/jira/browse/SPARK-23130 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 1.6.3, 2.1.0, 2.2.0 > Environment: * Hadoop distributions: HDP 2.5 - 2.6.3.0 > * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 >Reporter: Sean Roberts >Priority: Major > > Spark Thrift is not cleaning up /tmp for files & directories named like: > /tmp/hive/*.pipeout > /tmp/*_resources > There are such a large number that /tmp quickly runs out of inodes *causing > the partition to be unusable and many services to crash*. This is even true > when the only jobs submitted are routine service checks. > Used `strace` to show that Spark Thrift is responsible: > {code:java} > strace.out.118864:04:53:49 >
[jira] [Updated] (SPARK-23130) Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout)
[ https://issues.apache.org/jira/browse/SPARK-23130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts updated SPARK-23130: - Environment: * Spark versions: 1.6.3, 2.1.0, 2.2.0 * Hadoop distributions: HDP 2.5 - 2.6.3.0 * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 was: * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 * Spark versions: 1.6.3, 2.1.0, 2.2.0 * Hadoop distributions: HDP 2.5 - 2.6.3.0 > Spark Thrift does not clean-up temporary files (/tmp/*_resources and > /tmp/hive/*.pipeout) > - > > Key: SPARK-23130 > URL: https://issues.apache.org/jira/browse/SPARK-23130 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 1.6.3, 2.1.0, 2.2.0 > Environment: * Spark versions: 1.6.3, 2.1.0, 2.2.0 > * Hadoop distributions: HDP 2.5 - 2.6.3.0 > * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 >Reporter: Sean Roberts >Priority: Major > > Spark Thrift is not cleaning up /tmp for files & directories named like: > /tmp/hive/*.pipeout > /tmp/*_resources > There are such a large number that /tmp quickly runs out of inodes *causing > the partition to be unusable and many services to crash*. This is even true > when the only jobs submitted are routine service checks. > Used `strace` to show that Spark Thrift is responsible: > {code:java} > strace.out.118864:04:53:49 > open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", > O_RDWR|O_CREAT|O_EXCL, 0666) = 134 > strace.out.118864:04:53:49 > mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 > {code} > *Those files were left behind, even days later.* > ** > > Example files: > {code:java} > # stat > /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout > File: > ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ > Size: 0 Blocks: 0 IO Block: 4096 regular empty file > Device: fe09h/65033d Inode: 678 Links: 1 > Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:53:49.126777260 -0600 > Modify: 2017-12-19 04:53:49.126777260 -0600 > Change: 2017-12-19 04:53:49.126777260 -0600 > Birth: - > # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources > File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ > Size: 4096 Blocks: 8 IO Block: 4096 directory > Device: fe09h/65033d Inode: 668 Links: 2 > Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:57:38.458937635 -0600 > Modify: 2017-12-19 04:53:49.062777216 -0600 > Change: 2017-12-19 04:53:49.066777218 -0600 > Birth: - > {code} > Showing the large number: > {code:java} > # find /tmp/ -name '*_resources' | wc -l > 68340 > # find /tmp/hive -name "*.pipeout" | wc -l > 51837 > {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-23130) Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout)
[ https://issues.apache.org/jira/browse/SPARK-23130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts updated SPARK-23130: - Environment: * Hadoop distributions: HDP 2.5 - 2.6.3.0 * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 was: * Spark versions: 1.6.3, 2.1.0, 2.2.0 * Hadoop distributions: HDP 2.5 - 2.6.3.0 * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 > Spark Thrift does not clean-up temporary files (/tmp/*_resources and > /tmp/hive/*.pipeout) > - > > Key: SPARK-23130 > URL: https://issues.apache.org/jira/browse/SPARK-23130 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 1.6.3, 2.1.0, 2.2.0 > Environment: * Hadoop distributions: HDP 2.5 - 2.6.3.0 > * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 >Reporter: Sean Roberts >Priority: Major > > Spark Thrift is not cleaning up /tmp for files & directories named like: > /tmp/hive/*.pipeout > /tmp/*_resources > There are such a large number that /tmp quickly runs out of inodes *causing > the partition to be unusable and many services to crash*. This is even true > when the only jobs submitted are routine service checks. > Used `strace` to show that Spark Thrift is responsible: > {code:java} > strace.out.118864:04:53:49 > open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", > O_RDWR|O_CREAT|O_EXCL, 0666) = 134 > strace.out.118864:04:53:49 > mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 > {code} > *Those files were left behind, even days later.* > ** > > Example files: > {code:java} > # stat > /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout > File: > ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ > Size: 0 Blocks: 0 IO Block: 4096 regular empty file > Device: fe09h/65033d Inode: 678 Links: 1 > Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:53:49.126777260 -0600 > Modify: 2017-12-19 04:53:49.126777260 -0600 > Change: 2017-12-19 04:53:49.126777260 -0600 > Birth: - > # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources > File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ > Size: 4096 Blocks: 8 IO Block: 4096 directory > Device: fe09h/65033d Inode: 668 Links: 2 > Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) > Access: 2017-12-19 04:57:38.458937635 -0600 > Modify: 2017-12-19 04:53:49.062777216 -0600 > Change: 2017-12-19 04:53:49.066777218 -0600 > Birth: - > {code} > Showing the large number: > {code:java} > # find /tmp/ -name '*_resources' | wc -l > 68340 > # find /tmp/hive -name "*.pipeout" | wc -l > 51837 > {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Created] (SPARK-23130) Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout)
Sean Roberts created SPARK-23130: Summary: Spark Thrift does not clean-up temporary files (/tmp/*_resources and /tmp/hive/*.pipeout) Key: SPARK-23130 URL: https://issues.apache.org/jira/browse/SPARK-23130 Project: Spark Issue Type: Bug Components: SQL Affects Versions: 2.2.0, 2.1.0, 1.6.3 Environment: * OS: Seen on SLES12, RHEL 7.3 & RHEL 7.4 * Spark versions: 1.6.3, 2.1.0, 2.2.0 * Hadoop distributions: HDP 2.5 - 2.6.3.0 Reporter: Sean Roberts Spark Thrift is not cleaning up /tmp for files & directories named like: /tmp/hive/*.pipeout /tmp/*_resources There are such a large number that /tmp quickly runs out of inodes *causing the partition to be unusable and many services to crash*. This is even true when the only jobs submitted are routine service checks. Used `strace` to show that Spark Thrift is responsible: {code:java} strace.out.118864:04:53:49 open("/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout", O_RDWR|O_CREAT|O_EXCL, 0666) = 134 strace.out.118864:04:53:49 mkdir("/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources", 0777) = 0 {code} *Those files were left behind, even days later.* ** Example files: {code:java} # stat /tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout File: ‘/tmp/hive/55ad7fc1-f79a-4ad8-8e02-26bbeaa86bbc7288010135864174970.pipeout’ Size: 0 Blocks: 0 IO Block: 4096 regular empty file Device: fe09h/65033dInode: 678 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) Access: 2017-12-19 04:53:49.126777260 -0600 Modify: 2017-12-19 04:53:49.126777260 -0600 Change: 2017-12-19 04:53:49.126777260 -0600 Birth: - # stat /tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources File: ‘/tmp/b6dfbf9e-2f7c-4c25-95a1-73c44318ecf4_resources’ Size: 4096Blocks: 8 IO Block: 4096 directory Device: fe09h/65033dInode: 668 Links: 2 Access: (0700/drwx--) Uid: ( 1000/hive) Gid: ( 1002/ hadoop) Access: 2017-12-19 04:57:38.458937635 -0600 Modify: 2017-12-19 04:53:49.062777216 -0600 Change: 2017-12-19 04:53:49.066777218 -0600 Birth: - {code} Showing the large number: {code:java} # find /tmp/ -name '*_resources' | wc -l 68340 # find /tmp/hive -name "*.pipeout" | wc -l 51837 {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-23089) "Unable to create operation log session directory" when parent directory not present
[ https://issues.apache.org/jira/browse/SPARK-23089?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts updated SPARK-23089: - Environment: /usr/hdp/2.6.3.0-235/spark2/jars/spark-hive-thriftserver_2.11-2.2.0.2.6.3.0-235.jar $ cat /etc/redhat-release Red Hat Enterprise Linux Server release 7.4 (Maipo) $ ps aux|grep ^hive.*spark.*thrift hive 1468503 0.9 0.5 13319628 1411676 ?Sl Jan15 10:18 /usr/java/default/bin/java -Dhdp.version=2.6.3.0-235 -cp /usr/hdp/current/spark2-thriftserver/conf/:/usr/hdp/current/spark2-thriftserver/jars/*:/usr/hdp/current/hadoop-client/conf/ -Xmx2048m org.apache.spark.deploy.SparkSubmit --properties-file /usr/hdp/current/spark2-thriftserver/conf/spark-thrift-sparkconf.conf --class org.apache.spark.sql.hive.thriftserver.HiveThriftServer2 --name Thrift JDBC/ODBC Server spark-internal > "Unable to create operation log session directory" when parent directory not > present > > > Key: SPARK-23089 > URL: https://issues.apache.org/jira/browse/SPARK-23089 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 2.2.0 > Environment: > /usr/hdp/2.6.3.0-235/spark2/jars/spark-hive-thriftserver_2.11-2.2.0.2.6.3.0-235.jar > $ cat /etc/redhat-release > Red Hat Enterprise Linux Server release 7.4 (Maipo) > $ ps aux|grep ^hive.*spark.*thrift > hive 1468503 0.9 0.5 13319628 1411676 ?Sl Jan15 10:18 > /usr/java/default/bin/java -Dhdp.version=2.6.3.0-235 -cp > /usr/hdp/current/spark2-thriftserver/conf/:/usr/hdp/current/spark2-thriftserver/jars/*:/usr/hdp/current/hadoop-client/conf/ > -Xmx2048m org.apache.spark.deploy.SparkSubmit --properties-file > /usr/hdp/current/spark2-thriftserver/conf/spark-thrift-sparkconf.conf --class > org.apache.spark.sql.hive.thriftserver.HiveThriftServer2 --name Thrift > JDBC/ODBC Server spark-internal >Reporter: Sean Roberts >Priority: Minor > > When creating a session directory, Thrift should create the parent directory > _(i.e. /tmp/hive/operation_logs)_ if it is not present. > It's common for operators to clean-up old and empty directories in /tmp, or > to have tools (systemd-tmpfiles or tmpwatch) that do it automatically. > This was fixed in HIVE-12262 but not in Spark Thrift as seen by this: > {code}18/01/15 14:22:49 WARN HiveSessionImpl: Unable to create operation log > session directory: > /tmp/hive/operation_logs/683a6318-adc4-42c4-b665-11dad14d7ec7{code} > Resolved by manually creating /tmp/hive/operation_logs/ -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-23089) "Unable to create operation log session directory" when parent directory not present
[ https://issues.apache.org/jira/browse/SPARK-23089?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts updated SPARK-23089: - Description: When creating a session directory, Thrift should create the parent directory _(i.e. /tmp/hive/operation_logs)_ if it is not present. It's common for operators to clean-up old and empty directories in /tmp, or to have tools (systemd-tmpfiles or tmpwatch) that do it automatically. This was fixed in HIVE-12262 but not in Spark Thrift as seen by this: {code}18/01/15 14:22:49 WARN HiveSessionImpl: Unable to create operation log session directory: /tmp/hive/operation_logs/683a6318-adc4-42c4-b665-11dad14d7ec7{code} Resolved by manually creating /tmp/hive/operation_logs/ We are experiencing this with Spark Thrift provided by: /usr/hdp/2.6.3.0-235/spark2/jars/spark-hive-thriftserver_2.11-2.2.0.2.6.3.0-235.jar > "Unable to create operation log session directory" when parent directory not > present > > > Key: SPARK-23089 > URL: https://issues.apache.org/jira/browse/SPARK-23089 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 2.2.0 >Reporter: Sean Roberts >Priority: Minor > > When creating a session directory, Thrift should create the parent directory > _(i.e. /tmp/hive/operation_logs)_ if it is not present. > It's common for operators to clean-up old and empty directories in /tmp, or > to have tools (systemd-tmpfiles or tmpwatch) that do it automatically. > This was fixed in HIVE-12262 but not in Spark Thrift as seen by this: > {code}18/01/15 14:22:49 WARN HiveSessionImpl: Unable to create operation log > session directory: > /tmp/hive/operation_logs/683a6318-adc4-42c4-b665-11dad14d7ec7{code} > Resolved by manually creating /tmp/hive/operation_logs/ > We are experiencing this with Spark Thrift provided by: > /usr/hdp/2.6.3.0-235/spark2/jars/spark-hive-thriftserver_2.11-2.2.0.2.6.3.0-235.jar -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-23089) "Unable to create operation log session directory" when parent directory not present
[ https://issues.apache.org/jira/browse/SPARK-23089?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts updated SPARK-23089: - Environment: (was: When creating a session directory, Thrift should create the parent directory _(i.e. /tmp/hive/operation_logs)_ if it is not present. It's common for operators to clean-up old and empty directories in /tmp, or to have tools (systemd-tmpfiles or tmpwatch) that do it automatically. This was fixed in HIVE-12262 but not in Spark Thrift as seen by this: {code}18/01/15 14:22:49 WARN HiveSessionImpl: Unable to create operation log session directory: /tmp/hive/operation_logs/683a6318-adc4-42c4-b665-11dad14d7ec7{code} Resolved by manually creating /tmp/hive/operation_logs/ We are experiencing this with Spark Thrift provided by: /usr/hdp/2.6.3.0-235/spark2/jars/spark-hive-thriftserver_2.11-2.2.0.2.6.3.0-235.jar) > "Unable to create operation log session directory" when parent directory not > present > > > Key: SPARK-23089 > URL: https://issues.apache.org/jira/browse/SPARK-23089 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 2.2.0 >Reporter: Sean Roberts >Priority: Minor > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-23089) "Unable to create operation log session directory" when parent directory not present
[ https://issues.apache.org/jira/browse/SPARK-23089?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Sean Roberts updated SPARK-23089: - Description: When creating a session directory, Thrift should create the parent directory _(i.e. /tmp/hive/operation_logs)_ if it is not present. It's common for operators to clean-up old and empty directories in /tmp, or to have tools (systemd-tmpfiles or tmpwatch) that do it automatically. This was fixed in HIVE-12262 but not in Spark Thrift as seen by this: {code}18/01/15 14:22:49 WARN HiveSessionImpl: Unable to create operation log session directory: /tmp/hive/operation_logs/683a6318-adc4-42c4-b665-11dad14d7ec7{code} Resolved by manually creating /tmp/hive/operation_logs/ was: When creating a session directory, Thrift should create the parent directory _(i.e. /tmp/hive/operation_logs)_ if it is not present. It's common for operators to clean-up old and empty directories in /tmp, or to have tools (systemd-tmpfiles or tmpwatch) that do it automatically. This was fixed in HIVE-12262 but not in Spark Thrift as seen by this: {code}18/01/15 14:22:49 WARN HiveSessionImpl: Unable to create operation log session directory: /tmp/hive/operation_logs/683a6318-adc4-42c4-b665-11dad14d7ec7{code} Resolved by manually creating /tmp/hive/operation_logs/ We are experiencing this with Spark Thrift provided by: /usr/hdp/2.6.3.0-235/spark2/jars/spark-hive-thriftserver_2.11-2.2.0.2.6.3.0-235.jar > "Unable to create operation log session directory" when parent directory not > present > > > Key: SPARK-23089 > URL: https://issues.apache.org/jira/browse/SPARK-23089 > Project: Spark > Issue Type: Bug > Components: SQL >Affects Versions: 2.2.0 >Reporter: Sean Roberts >Priority: Minor > > When creating a session directory, Thrift should create the parent directory > _(i.e. /tmp/hive/operation_logs)_ if it is not present. > It's common for operators to clean-up old and empty directories in /tmp, or > to have tools (systemd-tmpfiles or tmpwatch) that do it automatically. > This was fixed in HIVE-12262 but not in Spark Thrift as seen by this: > {code}18/01/15 14:22:49 WARN HiveSessionImpl: Unable to create operation log > session directory: > /tmp/hive/operation_logs/683a6318-adc4-42c4-b665-11dad14d7ec7{code} > Resolved by manually creating /tmp/hive/operation_logs/ -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Created] (SPARK-23089) "Unable to create operation log session directory" when parent directory not present
Sean Roberts created SPARK-23089: Summary: "Unable to create operation log session directory" when parent directory not present Key: SPARK-23089 URL: https://issues.apache.org/jira/browse/SPARK-23089 Project: Spark Issue Type: Bug Components: SQL Affects Versions: 2.2.0 Environment: When creating a session directory, Thrift should create the parent directory _(i.e. /tmp/hive/operation_logs)_ if it is not present. It's common for operators to clean-up old and empty directories in /tmp, or to have tools (systemd-tmpfiles or tmpwatch) that do it automatically. This was fixed in HIVE-12262 but not in Spark Thrift as seen by this: {code}18/01/15 14:22:49 WARN HiveSessionImpl: Unable to create operation log session directory: /tmp/hive/operation_logs/683a6318-adc4-42c4-b665-11dad14d7ec7{code} Resolved by manually creating /tmp/hive/operation_logs/ We are experiencing this with Spark Thrift provided by: /usr/hdp/2.6.3.0-235/spark2/jars/spark-hive-thriftserver_2.11-2.2.0.2.6.3.0-235.jar Reporter: Sean Roberts -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org