Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 635611

Summary:	specific machine jobs haven't got higher priority than no machine specific ones
Product:	[Retired] Beaker	Reporter:	Vladimir Benes <vbenes>
Component:	scheduler	Assignee:	Bill Peck <bpeck>
Status:	CLOSED CURRENTRELEASE	QA Contact:
Severity:	medium	Docs Contact:
Priority:	low
Version:	0.7	CC:	bpeck, dcallagh, kbaker, mcsontos, pmuller, rmancy
Target Milestone:	---	Keywords:	FutureFeature
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Enhancement
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2010-10-28 02:02:12 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Vladimir Benes 2010-09-20 10:27:50 UTC

Description of problem:
when I need specific machine and there exist a bunch of jobs prior me in the queue it can take ages to get to the machine as other jobs which can be scheduled on other, currently not demanded machine, are scheduled on the one I need.

Version-Release number of selected component (if applicable):
Version - 0.5.57 

How reproducible:


Steps to Reproduce:
1.obtain a machine 
2.create new job for reserving that machine chosen above 
3.cancel the first job
  
Actual results:
for 99% somebody else will jump on that machine even in case he didn't specify the hostname

Expected results:
check if the machine is not explicitly needed by other jobs before scheduling

Additional info:

Comment 1 Marian Csontos 2010-09-20 11:35:19 UTC

More on the topic here:

  http://post-office.corp.redhat.com/archives/beaker-dev-list/2010-September/msg00052.html

I guess this is not something what would happen any time soon.

Comment 2 Petr Muller 2010-09-20 11:45:25 UTC

(In reply to comment #1) 
> I guess this is not something what would happen any time soon.

Please reconsider this part, because this makes hw-dependent testing a nightmare. And hw-dependent testing is nothing unusual. I spent last few weeks testing various instruction-set optimizations, and seeing a rare system grokking yet another generic job is very frustrating.

Comment 3 Vladimir Benes 2010-09-20 11:59:49 UTC

> Please reconsider this part, because this makes hw-dependent testing a
> nightmare. And hw-dependent testing is nothing unusual. I spent last few weeks
> testing various instruction-set optimizations, and seeing a rare system
> grokking yet another generic job is very frustrating.

yes, exactly, I am waiting for exact system half a day+ so the idea with priorities looks good to me.

Comment 5 Bill Peck 2010-09-20 14:07:47 UTC

This might be an easy fix if we automatically bump the priority of a recipe if it only matches one machine.