| *** mhen_ is now known as mhen | 02:30 | |
| *** mklejn_ is now known as mklejn | 08:35 | |
| npanda | I am trying to stack the VMs to one hypervisor until all the available CPU exhausts. I am trying to use | 13:19 |
|---|---|---|
| npanda | cpu_weight_multiplier as described at https://docs.openstack.org/nova/zed/configuration/config.html#filter_scheduler.weight_classes | 13:20 |
| npanda | I still can't achieve it. The VMs are getting spread across all hypervisors | 13:21 |
| npanda | I am using | 13:21 |
| npanda | [filter_scheduler] enabled_filters = MetricsFilter,AvailabilityZoneFilter,AggregateInstanceExtraSpecsFilter,DifferentHostFilter,SameHostFilter,ComputeFilter,ServerGroupAntiAffinityFilter,ServerGroupAffinityFilter,ImagePropertiesFilter,ComputeCapabilitiesFilter,NUMATopologyFilter | 13:21 |
| npanda | host_subset_size = 1 | 13:21 |
| npanda | weight_classes = nova.scheduler.weights.all_weighers | 13:22 |
| npanda | cpu_weight_multiplier = -1.0 | 13:22 |
| npanda | I wounder how do I enable this. <Note that this setting only affects scheduling if the CPUWeigher weigher is enabled.> | 13:25 |
| *** d34dh0r5- is now known as d34dh0r53 | 13:39 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!