|
89 | 89 | <!--See the condorProperties bean for examples of supported config options--> |
90 | 90 | </bean> |
91 | 91 |
|
92 | | - <!--This is an example of using docker for executing jobs. This is not mutually exclusive with the options above--> |
93 | | - <!--Your server must still execute condor_submit, and you will need to configure something along the lines of the configuration above for this--> |
94 | | - <!--Because this config uses DockerHTCondorExecutionEngineConfig, this means the engine will run within docker--> |
95 | | - <bean id="condorDockerProperties" class="org.labkey.cluster.pipeline.DockerHTCondorExecutionEngineConfig"> |
96 | | - <property name="location" value="clusterDocker" /> |
97 | | - |
98 | | - <!--Note: you will likely need to configure submitCommand, statusCommand, etc. as above--> |
99 | | - |
100 | | - <!--The path to docker on the cluster.--> |
101 | | - <!--<property name="remoteExecutable" value="docker" />--> |
102 | | - |
103 | | - <!--See the condorProperties bean for examples of supported config options. All of the same options are supported here--> |
104 | | - <!--In addition to those, the following additional options are supported:--> |
105 | | - <!--If using ActiveMQ, this should be a URL, relative to the cluster nodes, of your ActiveMQ server--> |
106 | | - <!--<property name="activeMqHost" value=""/>--> |
107 | | - <!--Your remote nodes will need a pipelineConfig.xml, comparable to this one. This specifies the folder holding this file--> |
108 | | - <!--<property name="configDir" value=""/>--> |
109 | | - <!--You do not necessarily need to base of this image, but it will have matching LabKey code. XX.X should match your server's version--> |
110 | | - <!--<property name="dockerImageName" value="bbimber/discvr-seq:XX.X"/>--> |
111 | | - </bean> |
112 | | - |
113 | 92 | <!--Note: A another route that might be useful for some clusters is to wrap java with a simple bash script. This can be useful if you need to set aspects of the environment--> |
114 | 93 | <!--and do not have enough control over the cluster to do this directly. For example, we needed to set 'umask 0002' on one cluster, and this was the best route we found--> |
115 | 94 | <!--The rough pattern would involve changing the remoteExecutable to point to your wrapper: --> |
|
0 commit comments