Opened 13 years ago
Last modified 10 years ago
#1172 new task
IZ231: Need SLO test
Reported by: | crei | Owned by: | |
---|---|---|---|
Priority: | high | Milestone: | |
Component: | testsuite | Version: | current |
Severity: | Keywords: | Sun hedeby | |
Cc: |
Description
[Imported from gridengine issuezilla http://gridengine.sunsource.net/issues/show_bug.cgi?id=231]
Issue #: 231 Platform: Sun Reporter: crei (crei) Component: testsuite OS: All Subcomponent: hedeby Version: current CC: None defined Status: NEW Priority: P2 Resolution: Issue type: TASK Target milestone: milestone 1 Assigned to: crei (crei) QA Contact: crei URL: * Summary: Need SLO test Status whiteboard: Attachments: Issue 231 blocks: Votes for issue 231: Vote for this issue Opened: Wed Aug 6 05:24:00 -0700 2008 ------------------------ slo/FixedUsage All ge services should have only one SLO (a FixedUsageSLO) modify urgency of FixedUsageSLO of ge service check the usage of all resources Lower the urgency of the FixedUsageSLO of ge service (lower then the urgency of the PermanentRequestSLO of the spare pool) Test that all resource goes into spare pool. Increase the urgency of the FixedUsageSLO of the ge service (higher the urgency of the PermanentRequestSLO of the spare pool) resource must stay in spare_pool because FixedUsageSLO does not produce any need slo/MinResource Define a MinResourceSLO which will need all host resources with arch A of the system for service GE1. Define a MinResourceSLO which will need all host resources with arch B of the system for service GE2. Check that the resources moves to the correct service Check the usage of the assigned resources. Reduce the min number of resource for both SLOs. Check that the resource goes back into spare pool. Increase the number of resource to a specific number. Check that the service will not get more then the specified number of resources Modify the resource filter that it does not SLO does not count resource which are in ERROR state. Kill the execds of some assigned host resources check that additional resources are assigned to the service. Cleanup slo/MaxPendingJobs Move all resources into spare_pool (not the static resources) disable all qinstances in the static resource setup MaxPendingJobsSLO (not more then one pending job) submit a sleeper job check that one resource is assigned to service wait until jobs finshed check that the resource goes back into spare_pool setup MaxPendingJobsSLO with a specific job filter (e.g. arch=lx24-amd64) submit a matching job resource must be assigned to service jobs run resource goes back into spare pool submit a not matching job no resource request must be produced no resource must be assigned jobs must not start delete the job ------- Additional comments from rhierlmeier Thu Feb 28 03:56:22 -0700 2008 ------- It's now as task ------- Additional comments from crei Fri Apr 4 01:38:31 -0700 2008 ------- Starting with slo/FixedUsage test ... ------- Additional comments from crei Wed Apr 9 03:47:40 -0700 2008 ------- slo/FixedUsage checked-in CR-2008-04-09-0: Enhancem.: o cleanup of logging and time out values o reset_hedeby(): add slo reset o new procedure: create_permanent_request_slo() o renamed set_hedeby_slos() to set_hedeby_slos_config() and finished implementation TODO's o new procedure: reset_default_slos() o Install test now also do setup slos (by calling reset_default_slos() after adding ge services and spare_pool resources) o Fixed order of cleanup commands in basic move test o Added new test: "slo/fixedUsage/check.62.exp" (Part 1/3 of hedeby issue #385) (Need SLO test) ------- Additional comments from crei Mon Apr 21 05:32:46 -0700 2008 ------- checked in part 2/3: slo/MinResource: CR-2008-04-18-0: Enhancem.: o Added new test: "slo/minResource/check.62.exp" (Part 2/3 of hedeby issue #385) (Need SLO test) o hedeby_check_default_resources() now also checks resource properties settings (install test) o cleanup of slo helper procedures and timeouts create_fixed_usage_slo() create_min_resource_slo() create_permanent_request_slo() o new procedures: create_resource_and_filter() get_service_slo_info() wait_for_service_slo_info() get_show_resource_request_info() o slo update interval for TS is now one minute o fixed host name resolving problems (TS always using short hostnames) ------- Additional comments from crei Wed Aug 6 05:25:40 -0700 2008 ------- increased priority
Note: See
TracTickets for help on using
tickets.