search

dask / dask-ec2

Start a cluster in EC2 for dask.distributed
106 stars 37 forks source link

dask distributed fails on workers #106

Closed mynameisvinn closed 6 years ago

mynameisvinn commented 6 years ago

starting with a fresh build, dask-ec2 distributed dask does what it's supposed to do with scheduler/master, but seems fails with workers.

this is the error message:

Installing scheduler
+---------+----------------------+-----------------+
| Node ID | # Successful actions | # Failed action |
+=========+======================+=================+
| node-0  | 16                   | 0               |
+---------+----------------------+-----------------+
Installing workers
+---------+----------------------+-----------------+
| Node ID | # Successful actions | # Failed action |
+=========+======================+=================+
| node-1  | 10                   | 6               |
| node-3  | 10                   | 6               |
| node-2  | 10                   | 6               |
+---------+----------------------+-----------------+
Failed states for 'node-1'
  supervisord | dworker-running | dworker | running: One or more requisite failed: dask.distributed.dask-install, dask.distributed.distributed-install, dask.distributed.worker.dworker-update-supervisor, dask.distributed.update-pandas, dask.distributed.worker.dworker.conf
  pip | distributed-install | distributed | installed: An importable Python 2 pip module is required but could not be found on your system. This usually means that the system's pip package is not installed properly.
  cmd | dworker-update-supervisor | /usr/bin/supervisorctl -c /etc/supervisor/supervisord.conf update && sleep 2 | wait: One or more requisite failed: dask.distributed.worker.dworker.conf
  pip | dask-install | dask | installed: An importable Python 2 pip module is required but could not be found on your system. This usually means that the system's pip package is not installed properly.
  cmd | update-pandas | /opt/anaconda//bin/conda update pandas | run: One or more requisite failed: dask.distributed.distributed-install
  file | dworker.conf | /etc/supervisor/conf.d//dworker.conf | managed: One or more requisite failed: dask.distributed.dask-install, dask.distributed.distributed-install, dask.distributed.update-pandas

i imagine it has to do with the fact that dworker/dscheduler has been deprecated to dask-worker/dask-scheduler, but am not 100% sure.

i have got it working and - if im not the only person running into this problem - i will submit a PR shortly.

jcrist commented 6 years ago

What version of dask-ec2 are you using? The current release doesn't use any of those commands. If you're not on the latest release, I suggest upgrading and trying again:

Latest release: pip install -U dask-ec2

Master branch: pip install -U git+https://github.com/dask/dask-ec2.git