Current basic test runs 7 VMs with 16GB RAM each this is overloading even high end laptops that we use Lets assign : 4G for engine and hypervisor 2G for the storage controllers reduce the hypervisors to 2
The problem seems to still be happening, it seems it tries to spin up a vm with 4G in the host (and of course, failing to)
Created attachment 1098161 [details] vdsm log
Relevant VDSM log entry: Thread-112::DEBUG::2015-11-24 03:43:36,194::vm::3313::vm.Vm::(_run) vmId=`8d11c321-4b35-48df-bd5e-751d0182456b`::<?xml version="1.0" encoding="utf-8"?> <domain type="kvm"> <name>vm1</name> <uuid>8d11c321-4b35-48df-bd5e-751d0182456b</uuid> <memory>4194304</memory> <currentMemory>4194304</currentMemory> (and I indeed can see in libvirt log that it failed running it).
Failed also on jenkins http://jenkins.ovirt.org/job/ovirt_3.5_system-tests/31/consoleFull
Sorry, that job fails for different reasons (though the lago log is the same)
I assume the bug is in: 147 memory=4 * GB, @ ovirt-system-tests/basic_suite_3.5/test-scenarios/004_basic_sanity.py
that looks like it yes