Skip to content

Commit

Permalink
yarn: use system wide install of worker script
Browse files Browse the repository at this point in the history
  • Loading branch information
aabadie committed Jun 5, 2017
1 parent 900abbb commit 7cb7e85
Show file tree
Hide file tree
Showing 4 changed files with 15 additions and 16 deletions.
2 changes: 2 additions & 0 deletions joblibhadoop/yarn/remoteworker.py → bin/joblib-yarn-worker
100644 → 100755
Original file line number Diff line number Diff line change
@@ -1,3 +1,5 @@
#!/usr/bin/env python

"""Remote worker script."""

import sys
Expand Down
7 changes: 3 additions & 4 deletions joblibhadoop/yarn/pool.py
Original file line number Diff line number Diff line change
Expand Up @@ -18,13 +18,12 @@ def __init__(self, processes=None, port=0, authkey=None):
self.stopping = False
self.knit = Knit(autodetect=True)

cmd = ('python remoteworker.py --host {} --port {} --key {}'
cmd = ('joblib-yarn-worker --host {} --port {} --key {}'
.format(socket.gethostname(),
self.server.address[1],
self.authkey))
self.app_id = self.knit.start(
cmd, num_containers=self._processes,
files=['joblibhadoop/yarn/remoteworker.py', ])
self.app_id = self.knit.start(cmd,
num_containers=self._processes)
self.thread = Thread(target=self._monitor_appid)
self.thread.deamon = True
self.thread.start()
Expand Down
14 changes: 5 additions & 9 deletions joblibhadoop/yarn/remotepool.py
Original file line number Diff line number Diff line change
Expand Up @@ -112,15 +112,11 @@ def _start_remote_worker(self, pid):
remote_worker = RemoteWorker(pid)
debug('starting remote worker %d', pid)

args = ['python', self.workerscript]
args.append('--host')
args.append(socket.gethostname())
args.append('--port')
args.append(str(self.server.address[1]))
args.append('--workerid')
args.append(str(pid))
args.append('--key')
args.append(self.authkey.decode())
args = [self.workerscript,
'--host', socket.gethostname(),
'--port', str(self.server.address[1]),
'--workerid', str(pid),
'--key', self.authkey.decode()]

remote_worker.proc = subprocess.Popen(args)
self._pool.append(remote_worker)
Expand Down
8 changes: 5 additions & 3 deletions setup.py
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
"""joblib-hadoop package installation module."""

from os.path import join as pjoin
from setuptools import setup, find_packages

if __name__ == '__main__':
Expand All @@ -14,9 +15,10 @@
license='BSD',
platforms='any',
packages=find_packages(),
scripts=[pjoin('bin', 'joblib-yarn-worker')],
install_requires=[
'joblib>=0.10',
'knit>=0.2'
'joblib>=0.10',
'knit>=0.2'
],
zip_safe=False,
)
)

0 comments on commit 7cb7e85

Please sign in to comment.