Description of problem: The following commands use runInernalQuery mechanism, which suspends the current transaction and opens another one, in order to load host nics (GetVdsInterfacesByVdsId query): * ActivateVdsCommand * AttachNetworkToClusterInternalCommand * HostSetupNetworksCommand * MaintenanceNumberOfVdssCommand * UpdateNetworkCommand * UpdateNetworkOnClusterCommand * UpdateVdsCommand Version-Release number of selected component (if applicable): How reproducible: 100% Actual results: 1. The running command 2. Redundant additional transaction is open 3. If the parent transaction flow updated host nics in DB, running the query in a separate transaction will return results that aren't include those updates. Expected results: The commands should use the current transaction fo loading host nics. Additional info:
Verified on: Red Hat Virtualization Manager Version: 4.1.2-0.1.el7 In order to verify it, I done the following steps: 1. Adding additional core (2 in total) on the engine VM to improve engine CPU time. 2. Setting up environment with 30 hosts and 50 hosts networks with VLAN attached to each of the hosts. 3. Setting up Power Management agent on each of the 50 hosts. 4. Changing host comment. 5. Editing existing host network and changing IP address settings. 6. Removing and adding host networks from specific host.