Reputation: 3181
I am facing a weird behavior with the fabric2 command module. Those commands work like a charm:
connect = Connection(host=h, user=u ,connect_kwargs={"password":p})
connect.run('mkdir temp_streaming')
connect.put(m, "temp_streaming/mapper.py")
connect.put(r, "temp_streaming/reducer.py")
But when it comes to this one:
input_path = input("Please choose the input of your program (on your HDFS File System) ")
output_path = input("Please choose a name for your output folder ")
main_command = str('yarn jar '+ jar_path+' -files mapper.py,reducer.py -mapper temp_streaming/mapper.py -reducer temp_streaming/reducer.py -input '+ input_path + ' -output '+ output_path)
connect.run(main_command)
I get this error I can't figure out:
bash: yarn: command not found
Traceback (most recent call last):
File "__main.py__", line 77, in <module>
main()
File "__main.py__", line 65, in main
RunMapReduce(mapper, reducer, jar_path)
File "__main.py__", line 46, in RunMapReduce
connect.run(main_command)
File "<decorator-gen-3>", line 2, in run
File "/usr/local/lib/python3.5/dist-packages/fabric2/connection.py", line 30, in opens
return method(self, *args, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/fabric2/connection.py", line 586, in run
return self._run(runner, command, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/invoke/context.py", line 100, in _run
return runner.run(command, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/invoke/runners.py", line 268, in run
return self._run_body(command, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/invoke/runners.py", line 401, in _run_body
raise UnexpectedExit(result)
invoke.exceptions.UnexpectedExit: Encountered a bad command exit code!
Command: 'yarn jar /home/hadoop/hadoop/share/hadoop/tools/lib/hadoop-streaming-2.8.4.jar -files mapper.py,reducer.py -mapper temp_streaming/mapper.py -reducer temp_streaming/reducer.py -input /books/kafka_metamorphosis.txt -output /test'
Exit code: 127
Stdout: already printed
Stderr: already printed
It acts as if the distant server did not know the yarn command even though this bash command works when I directly log on the server with ssh and run it. It seems to me that it is related with this library but I couldn't solve my problem with the documentation.
Upvotes: 3
Views: 3756
Reputation: 472
I faced this problem too: turns out that the SSH run by Fabric is a non-interactive shell.
The $PATH
variable which contains the path to your yarn
bin is set in /etc/profile
or ~/.bash_profile
, or ~/.bashrc
, none of which get sourced.
Here is the link, along with some solutions.
Personally, I am not a fan of playing with those files. What I just did (prior to stumbling upon this link) is I used the full paths of the binaries : /usr/bin/yarn [the rest of your command]
Upvotes: 3
Reputation: 11
I somehow managed to get it to work with the keyword argument warn=True
like so:
result = connection.run(command, warn=True)
This option was not obvious on the documentation, needed to search an upgrade guide to find it. Although this seems not to be an issue, I would strongly suggest an improvement in Fabric2 documentation, as this is I reckon an immensely likely scenario.
Upvotes: 1