Is it possible to create a queue system for separate jobs. I have a module which roughly consumes all CPUs on our machine and if another user decides to run the same module with other data it should wait till all other runs are finished to prevent overloading. Is there a setting for this? Jasper Koehorst Wageningen UR
You have two options: (1) set the number of workers in the LocalJobRunner; (2) Galaxy can use many different computing clusters, all of which have queuing systems that you can control. See this page for more details on both options: http://wiki.galaxyproject.org/Admin/Config/Performance/Cluster#Local Good luck, J. On Jul 10, 2013, at 7:18 AM, Jasper Jan Koehorst wrote:
Is it possible to create a queue system for separate jobs.
I have a module which roughly consumes all CPUs on our machine and if another user decides to run the same module with other data it should wait till all other runs are finished to prevent overloading.
Is there a setting for this?
Jasper Koehorst Wageningen UR ___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
I tried the following in the xml: (it is modified that it is not a valid xml as I cannot deposit everything here yet….) the following somehow ignores the amount of jobs as when I run 2 the second one runs immediately after the first one in the history panel and in the terminal top I can see it as well. <tool id="X" name="name here" version="0.1"> <description>Annotation</description> <command interpreter="python">XXX.py '' '$' '$' '-' '$' '$' '$' '$' '$' '$' '$' '$' '$' '$' '$' '$'</command> <inputs> <param name="input" type="data" format="txt,html" label="File "/> <param name="x" type="select" label="x selection"> <option value="4">=4)</option> <option value="15">==15)</option> </param> <param name="gram" type="select" label="Gram positive/negative selection"> <option value="0">Unknown/Both</option> <option value="+"> Positive</option> <option value="-"> Negative</option> </param> <param name="ALL" type="boolean" truevalue="-all" falsevalue="" label="All options enabled"/> <param ... more... </inputs> <outputs> <data format="txt" name="outtxt" label="$input.name"/> <data format="html" name="outhtml" label="$input.name"/> </outputs> <plugins> <plugin id="local" type="runner" load="galaxy.jobs.runners.local:LocalJobRunner" workers="1"/> </plugins> <destinations> <destination id="local" runner="local"/> </destinations> </tool> On Jul 10, 2013, at 3:18 PM, Jeremy Goecks <jeremy.goecks@emory.edu> wrote:
You have two options:
(1) set the number of workers in the LocalJobRunner;
(2) Galaxy can use many different computing clusters, all of which have queuing systems that you can control.
See this page for more details on both options:
http://wiki.galaxyproject.org/Admin/Config/Performance/Cluster#Local
Good luck, J.
On Jul 10, 2013, at 7:18 AM, Jasper Jan Koehorst wrote:
Is it possible to create a queue system for separate jobs.
I have a module which roughly consumes all CPUs on our machine and if another user decides to run the same module with other data it should wait till all other runs are finished to prevent overloading.
Is there a setting for this?
Jasper Koehorst Wageningen UR ___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
participants (2)
-
Jasper Jan Koehorst -
Jeremy Goecks