0

Basically have a problem that has come up before: The same problem described here for example.

I get a connect error if I try to work a pig script.

Looking into it, indeed the history server is not running. Seemed easy enough to fix. But then... It was not.

Sadly, I tried to start the history server in both ways I am aware of:

1) Go into the hadoop sbin and work with:

   mr-jobhistory-daemon.sh start historyserver

Which results in:

  bash: mr-jobhistory-daemon.sh: command not found

Even though when I do an ls of the directory mr-jobhistory-daemon.sh is indeed listed.

2) simply do:

  sudo service hadoop-mapreduce-historyserver start

Which will at least don't throw a "not found" error. But instead:

  Failed to start Hadoop historyserver. Return value: 1      [FAILED]

With the logfile linked sadly turning out to be completely empty.

Hence: Same problem, but the solutions known to me seem not to work.

I am pretty sure I am overlooking something - as I am pretty sure I have had the same problem with a different setup a year ago. But I fail to remember how I solved it back then.

I am running the newest version of hadoop (2.7.2) on a centos VM with CentOS 6.5 on VMWare player.

Community
  • 1
  • 1
Nelnel
  • 61
  • 1
  • 5
  • try sh mr-jobhistory-daemon.sh start historyserver or bash mr-jobhistory-daemon.sh start historyserver or ./mr-jobhistory-daemon.sh start historyserver, did you get something different than 'bash: mr-jobhistory-daemon.sh: command not found' doing this ? – ar-ms May 26 '16 at 19:22
  • @Koffee Indeed this actually has some effect. I guess I know the problem, just not sure what to do to fix it. It says "cannot create dir", "missing operant after 'tom'" and then /bin/mapred no such file or directory. I would guess it is the later making the biggest problem. But I am confused why it is not there. – Nelnel May 28 '16 at 10:59

2 Answers2

0

Use this instead (./mr-jobhistory-daemon.sh start historyserver ) and verify it by running jps command from terminal!

0

I am using CentOS 6.2 and doing this worked for me

  • edit the .bashrc, using the command: sudo gedit ~/.bashrc

  • and add the path to your hadoop sbin: export PATH=$PATH:$/<user>/.../hadoop/hadoop-2.6.0/sbin

  • now close the terminal and reopen it.

  • run the command: source ~/.bashrc

  • now try running the history server: mr-jobhistory-daemon.sh start historyserver