Page MenuHomePhabricator

Problems on deployment-hadoop-test-1
Closed, ResolvedPublic

Description

Puppet is failing on deployment-hadoop-test-1 due to hdfs issues:

Error: /Stage[main]/Cdh::Hadoop::Namenode::Primary/Cdh::Hadoop::Directory[/tmp]/Kerberos::Exec[cdh::hadoop::directory /tmp]/Exec[cdh::hadoop::directory /tmp]/unless: Check "/usr/bin/hdfs dfs -test -e /tmp" exceeded timeout
Error: /Stage[main]/Cdh::Hadoop::Namenode::Primary/Cdh::Hadoop::Directory[/user]/Kerberos::Exec[cdh::hadoop::directory /user]/Exec[cdh::hadoop::directory /user]/unless: Check "/usr/bin/hdfs dfs -test -e /user" exceeded timeout
Error: /Stage[main]/Cdh::Hadoop::Namenode::Primary/Cdh::Hadoop::Directory[/user/hdfs]/Kerberos::Exec[cdh::hadoop::directory /user/hdfs]/Exec[cdh::hadoop::directory /user/hdfs]/unless: Check "/usr/bin/hdfs dfs -test -e /user/hdfs" exceeded timeout
Error: /Stage[main]/Cdh::Hadoop::Namenode::Primary/Cdh::Hadoop::Directory[/var]/Kerberos::Exec[cdh::hadoop::directory /var]/Exec[cdh::hadoop::directory /var]/unless: Check "/usr/bin/hdfs dfs -test -e /var" exceeded timeout
Error: /Stage[main]/Cdh::Hadoop::Namenode::Primary/Cdh::Hadoop::Directory[/var/lib]/Kerberos::Exec[cdh::hadoop::directory /var/lib]/Exec[cdh::hadoop::directory /var/lib]/unless: Check "/usr/bin/hdfs dfs -test -e /var/lib" exceeded timeout
Error: /Stage[main]/Cdh::Hadoop::Namenode::Primary/Cdh::Hadoop::Directory[/var/log]/Kerberos::Exec[cdh::hadoop::directory /var/log]/Exec[cdh::hadoop::directory /var/log]/unless: Check "/usr/bin/hdfs dfs -test -e /var/log" exceeded timeout
Error: /Stage[main]/Cdh::Hadoop::Resourcemanager/Cdh::Hadoop::Directory[/var/log/hadoop-yarn]/Kerberos::Exec[cdh::hadoop::directory /var/log/hadoop-yarn]/Exec[cdh::hadoop::directory /var/log/hadoop-yarn]/unless: Check "/usr/bin/hdfs dfs -test -e /var/log/hadoop-yarn" exceeded timeout

I don't know anything about debugging HDFS issues.
There are two home directories on this VM (other than mine) - @fgiunchedi (last login May 2019) and @Ottomata (last login June 2019). Is this instance still needed?
Puppet is fine on deployment-hadoop-test-[23]

Event Timeline

Ottomata claimed this task.

I don't think we need any of them. Deleted all 3.