| Package | Description |
|---|---|
| org.apache.uima.ducc.rm.scheduler |
| Modifier and Type | Field and Description |
|---|---|
protected ResourceClass |
RmJob.resource_class |
| Modifier and Type | Method and Description |
|---|---|
ResourceClass |
RmJob.getResourceClass() |
ResourceClass |
IRmJob.getResourceClass() |
ResourceClass |
Scheduler.getResourceClass(String name) |
ResourceClass |
ISchedulerMain.getResourceClass(String name) |
| Modifier and Type | Method and Description |
|---|---|
protected List<ResourceClass> |
NodepoolScheduler.traverseNodepoolsForCounts(org.apache.uima.ducc.rm.scheduler.NodePool np,
List<ResourceClass> eligible)
Do a depth-first traversal of the nodepool calculating counts for all the jobs in the nodepool and its children.
|
| Modifier and Type | Method and Description |
|---|---|
protected boolean |
NodepoolScheduler.countUserShares(ResourceClass rc)
Count out shares for only the jobs in the ResouceClasses here, and only from the given
nodepool.
|
void |
ResourceClass.initWantedByOrder(ResourceClass unused) |
void |
RmJob.initWantedByOrder(ResourceClass unused) |
void |
User.initWantedByOrder(ResourceClass rc) |
void |
RmJob.setResourceClass(ResourceClass cl)
Save ref to the class we are in, and init class-based structures.
|
void |
IRmJob.setResourceClass(ResourceClass cl)
Save ref to the class we are in, and init class-based structures.
|
| Modifier and Type | Method and Description |
|---|---|
protected void |
NodepoolScheduler.countClassShares(org.apache.uima.ducc.rm.scheduler.NodePool np,
List<ResourceClass> rcs)
Count out shares for only the jobs in the ResouceClasses here, and only from the given
nodepool.
|
protected void |
NodepoolScheduler.expandNeedyJobs(org.apache.uima.ducc.rm.scheduler.NodePool np,
ArrayList<ResourceClass> rcs)
Expand jobs from the needy list ahead of other jobs, because we likely stole the
shares from older jobs which would otherwise get priority.
|
protected void |
NodepoolScheduler.findHowMuch(ArrayList<ResourceClass> rcs)
The second stage - work up all counts globally
|
protected void |
NodepoolScheduler.findWhatOf(ArrayList<ResourceClass> rcs)
The third stage - work up all counts globally
|
protected boolean |
NodepoolScheduler.jobInClass(ArrayList<ResourceClass> rcs,
IRmJob j)
Is the job submitted to one of these classes?
|
void |
IScheduler.setClasses(Map<ResourceClass,ResourceClass> classes) |
void |
IScheduler.setClasses(Map<ResourceClass,ResourceClass> classes) |
void |
NodepoolScheduler.setClasses(Map<ResourceClass,ResourceClass> prclasses) |
void |
NodepoolScheduler.setClasses(Map<ResourceClass,ResourceClass> prclasses) |
protected List<ResourceClass> |
NodepoolScheduler.traverseNodepoolsForCounts(org.apache.uima.ducc.rm.scheduler.NodePool np,
List<ResourceClass> eligible)
Do a depth-first traversal of the nodepool calculating counts for all the jobs in the nodepool and its children.
|
protected void |
NodepoolScheduler.traverseNodepoolsForExpansion(org.apache.uima.ducc.rm.scheduler.NodePool np,
ArrayList<ResourceClass> rcs) |
protected void |
NodepoolScheduler.updateNodepools(org.apache.uima.ducc.rm.scheduler.NodePool np,
ArrayList<ResourceClass> rcs) |
protected void |
NodepoolScheduler.whatOfFairShare(ArrayList<ResourceClass> rcs)
Counts are established.
|
protected void |
NodepoolScheduler.whatOfFixedShare(ArrayList<ResourceClass> rcs)
If there are free shares of the right order just assign them.
|
Copyright © 2012–2015 The Apache Software Foundation. All rights reserved.