dispy3-3.4: question about dispynode

Help
jjremus
2012-07-25
2012-10-08
  • jjremus
    jjremus
    2012-07-25

    I am still working to get dispy set up on my cluster, so it is possible that
    the error below is my fault, but I thought it might be worthwhile to check
    with you. This is a networked node on the LAN. It looks like it connects
    properly and receives the job, but then tries to send the results to localhost
    rather than back to the head node. Am I missing something? I've included
    debugging output from the node worker below:

    jremus@peteandkenny:~/dispy_tutorial/dispy-3.4$ python3 dispynode.py -d -i
    192.168.0.100 --scheduler_node=192.168.0.120
    2012-07-24 23:25:19,782 - asyncoro - poller: epoll
    2012-07-24 23:25:19,783 - dispynode - auth_code for 192.168.0.100:
    b'e9c509c5fe56e699f199a6d6437ccec85ef303bd'
    2012-07-24 23:25:19,783 - dispynode - serving 8 cpus at 192.168.0.100:51348
    2012-07-24 23:25:19,783 - dispynode - tcp server at 192.168.0.100:51348
    2012-07-24 23:25:37,736 - dispynode - dest_path for "computeJJ":
    /tmp/dispy/67bac0a3127f5d36
    2012-07-24 23:25:37,737 - dispynode - xfer_files given:
    2012-07-24 23:25:50,921 - dispynode - New job id 1 from 192.168.0.120
    2012-07-24 23:25:52,929 - dispynode - Sending result for job 1 (10) to
    ('127.0.1.1', 51347)
    2012-07-24 23:25:52,930 - dispynode - Couldn't send results for 1 to
    ('127.0.1.1', 51347)
    2012-07-24 23:25:52,930 - dispynode - storing results for job 1
    2012-07-24 23:26:01,056 - dispynode - New job id 2 from 192.168.0.120
    2012-07-24 23:26:14,473 - dispynode - New job id 3 from 192.168.0.120
    2012-07-24 23:26:17,482 - dispynode - Sending result for job 3 (10) to
    ('127.0.1.1', 51347)
    2012-07-24 23:26:17,482 - dispynode - Couldn't send results for 3 to
    ('127.0.1.1', 51347)
    2012-07-24 23:26:17,483 - dispynode - storing results for job 3
    2012-07-24 23:26:17,815 - dispynode - New job id 4 from 192.168.0.120
    2012-07-24 23:26:18,821 - dispynode - Sending result for job 4 (10) to
    ('127.0.1.1', 51347)
    2012-07-24 23:26:18,822 - dispynode - Couldn't send results for 4 to
    ('127.0.1.1', 51347)
    2012-07-24 23:26:18,822 - dispynode - storing results for job 4
    2012-07-24 23:26:24,085 - dispynode - Sending result for job 2 (10) to
    ('127.0.1.1', 51347)
    2012-07-24 23:26:24,086 - dispynode - Couldn't send results for 2 to
    ('127.0.1.1', 51347)
    2012-07-24 23:26:24,086 - dispynode - storing results for job 2

     
  • Sorry for late reply - didn't notice this till now. It looks like default IP address at client is 127.0.0.1; if you set 'ip_addr' argument to JobCluster method (to 192.168.0.120) it should work.