The management network, named ovirtmgmt, is created during host bootstrap. It consists of a bridge device, connected to the network device that was used to communicate with Engine (nic, bonding or vlan). It inherits its ip settings from the latter device. Creating the management network is the most fragile, error-prone step of bootstrap. Currently it always creates a bridged network (even if the DC requires a non-bridged ovirtmgmt), it knows nothing about the defined MTU for ovirtmgmt, it uses ping to guess on top of which device to build (and thus requires Vdsm-to-Engine reverse connectivity), and is the sole remaining user of the addNetwork/vdsm-store-net-conf scripts. Bootstrap would avoid creating a management network. Instead, after bootstrapping a host, Engine would send a getVdsCaps probe to the installed host, receiving a complete picture of the network configuration on the host. Among this picture is the device that holds the host's management IP address. Most of the work lies in Engine, where the output of getVdsCaps should be parsed, and a setupNetworks command should be transmitted after a new host is added to the data center. A preliminary list of actions when adding a host is: 1. start host deployment with ODEPLOY/forceReboot = False VDSM/managementBridgeName undefined 2. after otopi finishes, start to poll host with getVdsCaps. If timeout expires, fail host deployment. 3. define ovirtmgmt on host - if already defined, declare success. - acquire lastClientInterface and devise network definition for ovirtmgmt. - send setupNetworks with the new network definition. - on success, send setSafeNetConfig. On failure show an event to the user. the host would be left unoperational, and may need manual network configuration. 4. if the user requested post-installation reboot, fence the newly-added host. http://lists.ovirt.org/pipermail/arch/2012-December/001101.html
*** Bug 917724 has been marked as a duplicate of this bug. ***
Hello, I see this merge upstream. I was not CCed on this, just to make sure... this patch should have deployed a host without creating the bridge, right? How was it tested? As far as mpavlik sees engine still passes the management bridge name to host-deploy. Thanks, Alon
(In reply to comment #3) > Hello, > > I see this merge upstream. > > I was not CCed on this, just to make sure... this patch should have deployed > a host without creating the bridge, right? > I missed a patch to the VdsDesploy which removes the need of this environment value. > How was it tested? > Since the entire logic was implemented as part of the host activation flow, it was tested via the activating the host and not by its installation. > As far as mpavlik sees engine still passes the management bridge name to > host-deploy. > Posted a complementary patch for the host installation part: http://gerrit.ovirt.org/14230 > Thanks, > Alon
I still need to complete the packaging.
Trying to rephrase the doc text a bit...
:) Thanks!
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHSA-2014-0038.html