[jira] [Created] (AMBARI-14497) NFS Gateway fails to start with /tmp/.hdfs-nfs warning in logs

["Andrew Onischuk (JIRA)" <ji...@apache.org>]

Andrew Onischuk created AMBARI-14497:
----------------------------------------

             Summary: NFS Gateway fails to start with /tmp/.hdfs-nfs warning in logs
                 Key: AMBARI-14497
                 URL: https://issues.apache.org/jira/browse/AMBARI-14497
             Project: Ambari
          Issue Type: Bug
            Reporter: Andrew Onischuk
            Assignee: Andrew Onischuk
             Fix For: 2.4.0
         Attachments: AMBARI-14497.patch

Steps:

  1. Before deploy were created log and PID dirs with root owners, root group and 000 permissions.
  2. Deployed clusters via blueprints.
  3. Removed cluster (reset + hostcleanup).
  4. Were recreated log and PID dirs with root owners, root group and 000 permissions.
  5. Deploy cluster via UI.

Result: NFS Gateway can not to be started:  
This looks like hdfs issue:

    
    
    Execution of 'ambari-sudo.sh  -H -E /usr/hdp/current/hadoop-client/sbin/hadoop-daemon.sh --config /usr/hdp/current/hadoop-client/conf start nfs3' returned 1. starting nfs3, logging to /grid/0/log/hadoop/root/hadoop-hdfs-nfs3-bug49661-5.out
    

Nfs gateway failed to start with this the following in logs:

    
    
    bug49661-5:~ # tail -f /grid/0/log/hadoop/root/hadoop-hdfs-nfs3-bug49661-5.log
    2015-12-16 17:28:45,064 INFO  impl.MetricsSystemImpl (MetricsSystemImpl.java:startTimer(377)) - Scheduled snapshot period at 60 second(s).
    2015-12-16 17:28:45,064 INFO  impl.MetricsSystemImpl (MetricsSystemImpl.java:start(192)) - Nfs3 metrics system started
    2015-12-16 17:28:45,086 INFO  oncrpc.RpcProgram (RpcProgram.java:<init>(84)) - Will accept client connections from unprivileged ports
    2015-12-16 17:28:45,094 INFO  security.ShellBasedIdMapping (ShellBasedIdMapping.java:updateStaticMapping(322)) - Not doing static UID/GID mapping because '/etc/nfs.map' does not exist.
    2015-12-16 17:28:45,105 INFO  nfs3.WriteManager (WriteManager.java:<init>(92)) - Stream timeout is 600000ms.
    2015-12-16 17:28:45,105 INFO  nfs3.WriteManager (WriteManager.java:<init>(100)) - Maximum open streams is 256
    2015-12-16 17:28:45,106 INFO  nfs3.OpenFileCtxCache (OpenFileCtxCache.java:<init>(54)) - Maximum open streams is 256
    2015-12-16 17:28:45,338 INFO  nfs3.RpcProgramNfs3 (RpcProgramNfs3.java:<init>(205)) - Configured HDFS superuser is 
    2015-12-16 17:28:45,339 INFO  nfs3.RpcProgramNfs3 (RpcProgramNfs3.java:clearDirectory(231)) - Delete current dump directory /tmp/.hdfs-nfs
    2015-12-16 17:28:45,343 WARN  fs.FileUtil (FileUtil.java:deleteImpl(187)) - Failed to delete file or dir [/tmp/.hdfs-nfs]: it still exists.
    

I checked and seems like /tmp/.hdfs-nfs was an existing folder, which I could
delete without problems (as root - since we run nfs gateway as root). Seems
like nfs issue, while it's no able to delete that folder for some reason.

    
    
    drwxr-xr-x 2      2815 hadoop        4096 Dec 16 13:40 .hdfs-nfs
    





--
This message was sent by Atlassian JIRA
(v6.3.4#6332)