Re: [galaxy-user] Galaxy CloudMan - Nodes can't make their own qsub calls?

2012-09-15 Thread Enis Afgan
Added: worker nodes are now submit hosts as well:
https://bitbucket.org/galaxy/cloudman/changeset/b4c66783b42fb559b35b689363b34826d00ced34

This will be generally available with the next release of CloudMan is made
(however, there is no immediate plans for doing so as quite a bit of other
details need to ironed out before the code is stable).



On Sat, Sep 15, 2012 at 4:36 AM, greg margeem...@gmail.com wrote:

 Hey all,

 Here's the one-liner I came up with for my users to run until this is
 changed in Galaxy:

 $ qconf -sconfl | xargs -l sudo -E -u sgeadmin
 /opt/sge/bin/lx24-amd64/qconf -as

 (in case anyone is interested)

 -Greg

 On Fri, Sep 14, 2012 at 2:27 PM, greg margeem...@gmail.com wrote:
  Spoke too soon!
 
   $ sudo -E -u sgeadmin /opt/sge/bin/lx24-amd64/qconf -as
  ip-10-29-213-80.ec2.internal
 
  worked fine.  -E tells sudo to use the users environment.
 
  -Greg
 
 
  On Fri, Sep 14, 2012 at 2:24 PM, greg margeem...@gmail.com wrote:
  Ok, here's what I tried:
 
  $ sudo -u sgeadmin qconf -as ip-10-29-213-80.ec2.internal
  sudo: qconf: command not found
 
  So then I tried:
 
  $ which qconf
  /opt/sge/bin/lx24-amd64/qconf
  $ sudo -u sgeadmin /opt/sge/bin/lx24-amd64/qconf -as
  ip-10-29-213-80.ec2.internal
  critical error: Please set the environment variable SGE_ROOT.
 
  So then I tried:
 
  $ echo $SGE_ROOT
  /opt/sge
  $ sudo -u sgeadmin export SGE_ROOT=/opt/sge
  sudo: export: command not found
 
 
  Any ideas?  I don't have the greatest understanding of the export
 command.
 
  Thanks again,
 
  Greg
 
 
  On Fri, Sep 14, 2012 at 8:23 AM, greg margeem...@gmail.com wrote:
  Thanks James.  I'll give that a try.  That would be excellent if you
  could make work nodes submit hosts by default in a future release.
  How would I get that on the plans?  Submit a ticket somewhere?
 
  I guess in the meantime I could write a python script that runs qconf
  -sconfl to list all the hosts and for any that aren't listed when
  running qconf -ss, it would run qconf -as hostname on them.
 
  -Greg
 
  On Thu, Sep 13, 2012 at 5:00 PM, James Taylor ja...@jamestaylor.org
 wrote:
  You probably need to sudo to sgeadmin (cloudman guys correct me if you
  have this setup differently).
 
  I don't see any reason not to make worker nodes submit hosts by
  default in a future cloudman release.
 
  -- jt
 
 
  On Thu, Sep 13, 2012 at 4:17 PM, greg margeem...@gmail.com wrote:
  As a follow up I found a command that should add the new nodes as
  submit hosts and I tried to run it but I got this error:
 
  $ qconf -as ip-10-28-164-178.ec2.internal
  denied: ubuntu must be manager for this operation
 
  What does it mean by manager?  How would I run this command?
 
 
  I guess my preference is for Cloudman to do this automatically though
  so I'll be distributing this program to 3rd party users using the
  built-in cloudman sharing.  I can't rightly ask users to be running
  qconf.
 
  Thanks again,
 
  Greg
 
 
  On Thu, Sep 13, 2012 at 3:59 PM, greg margeem...@gmail.com wrote:
  Hi guys,
 
  I created a new Galaxy instance web launcher
  (https://biocloudcentral.herokuapp.com/launch) and then I ssh'd
 into
  the master node.
 
  I'm trying to run a Perl script that makes several qsub calls to
 other
  perl scripts.  Now the catch is that one of those perl scripts makes
  its own qsub calls.
 
  And I'm getting this error when it tries to do that:
 
  Unable to run job: denied: host ip-10-29-176-111.ec2.internal is
 no
  submit host.
 
 
  Somehow this works fine on other clusters I've run this code on.
  Any
  idea what could be going on?  Do I need to make all of the nodes
  submit hosts?
 
  Thanks a bunch!
 
  -Greg
  ___
  The Galaxy User list should be used for the discussion of
  Galaxy analysis and other features on the public server
  at usegalaxy.org.  Please keep all replies on the list by
  using reply all in your mail client.  For discussion of
  local Galaxy instances and the Galaxy source code, please
  use the Galaxy Development list:
 
http://lists.bx.psu.edu/listinfo/galaxy-dev
 
  To manage your subscriptions to this and other Galaxy lists,
  please use the interface at:
 
http://lists.bx.psu.edu/
 ___
 The Galaxy User list should be used for the discussion of
 Galaxy analysis and other features on the public server
 at usegalaxy.org.  Please keep all replies on the list by
 using reply all in your mail client.  For discussion of
 local Galaxy instances and the Galaxy source code, please
 use the Galaxy Development list:

   http://lists.bx.psu.edu/listinfo/galaxy-dev

 To manage your subscriptions to this and other Galaxy lists,
 please use the interface at:

   http://lists.bx.psu.edu/

___
The Galaxy User list should be used for the discussion of
Galaxy analysis and other features on the public server
at 

Re: [galaxy-user] Galaxy CloudMan - Nodes can't make their own qsub calls?

2012-09-13 Thread James Taylor
You probably need to sudo to sgeadmin (cloudman guys correct me if you
have this setup differently).

I don't see any reason not to make worker nodes submit hosts by
default in a future cloudman release.

-- jt


On Thu, Sep 13, 2012 at 4:17 PM, greg margeem...@gmail.com wrote:
 As a follow up I found a command that should add the new nodes as
 submit hosts and I tried to run it but I got this error:

 $ qconf -as ip-10-28-164-178.ec2.internal
 denied: ubuntu must be manager for this operation

 What does it mean by manager?  How would I run this command?


 I guess my preference is for Cloudman to do this automatically though
 so I'll be distributing this program to 3rd party users using the
 built-in cloudman sharing.  I can't rightly ask users to be running
 qconf.

 Thanks again,

 Greg


 On Thu, Sep 13, 2012 at 3:59 PM, greg margeem...@gmail.com wrote:
 Hi guys,

 I created a new Galaxy instance web launcher
 (https://biocloudcentral.herokuapp.com/launch) and then I ssh'd into
 the master node.

 I'm trying to run a Perl script that makes several qsub calls to other
 perl scripts.  Now the catch is that one of those perl scripts makes
 its own qsub calls.

 And I'm getting this error when it tries to do that:

 Unable to run job: denied: host ip-10-29-176-111.ec2.internal is no
 submit host.


 Somehow this works fine on other clusters I've run this code on.  Any
 idea what could be going on?  Do I need to make all of the nodes
 submit hosts?

 Thanks a bunch!

 -Greg
 ___
 The Galaxy User list should be used for the discussion of
 Galaxy analysis and other features on the public server
 at usegalaxy.org.  Please keep all replies on the list by
 using reply all in your mail client.  For discussion of
 local Galaxy instances and the Galaxy source code, please
 use the Galaxy Development list:

   http://lists.bx.psu.edu/listinfo/galaxy-dev

 To manage your subscriptions to this and other Galaxy lists,
 please use the interface at:

   http://lists.bx.psu.edu/
___
The Galaxy User list should be used for the discussion of
Galaxy analysis and other features on the public server
at usegalaxy.org.  Please keep all replies on the list by
using reply all in your mail client.  For discussion of
local Galaxy instances and the Galaxy source code, please
use the Galaxy Development list:

  http://lists.bx.psu.edu/listinfo/galaxy-dev

To manage your subscriptions to this and other Galaxy lists,
please use the interface at:

  http://lists.bx.psu.edu/