similar to: Patch supercedes previous patch...

Replaced the previous HostWrapper class with the new HostAgent class, which more closely resembles the style of the ProcessorsAgent class. Deleted the HostWrapper class entirely. Refactored the daemon entry point to reduce the number of methods invoked in order to start or stop the HostAgent instance. Moved the NICWrapper's setupQMFObject method from private to public so that it can be

Refactoring the previous patch ended up creating two deltas. This patch pushes them both together into a single commit.

This patch includes one small change: the Processors::get_load_average() method is now const since it does not change the object's state.

After some feedback from Slow, mainly about the load_average API being a method rather than an ongoing statistic. So I've converted the code over to instead update the load average statistic on a regular basis.

Created a new class to represent network interface devices, named NetworkDeviceAgent. Signed-off-by: Darryl L. Pierce <dpierce at redhat.com> --- src/Makefile.am | 9 +- src/host.cpp | 26 +++-- src/host.h | 4 +- src/linux_platform.cpp | 70 +++++++++++ src/linux_platform.h | 4 + src/networkdevice.cpp | 49 ++++++++ src/networkdevice.h |

This class has been replaced by the ProcessorsAgent. Signed-off-by: Darryl L. Pierce <dpierce at redhat.com> --- src/Makefile.am | 4 - src/cpu.cpp | 216 ------------------------------------------------------- src/cpu.h | 111 ---------------------------- src/host.cpp | 24 ------ src/host.h | 3 - src/schema.xml | 16 ---- 6 files changed, 0 insertions(+),

Stripped out the HAL support code and replaced with calls to udev. The remainder of the code to extract CPU details parses through the /proc/cpuinfo file since udev/sysfs will not return such information. Signed-off-by: Darryl L. Pierce <dpierce at redhat.com> --- configure.ac | 3 +- src/Makefile.am | 4 +- src/processors.cpp | 94

The previous patch, for whatever reason, fails to apply on top of tip. This patch is reimplemented and, hopefully, fixed.

This patch changes no functionality. It merely adds items to the ignore file, cleans up the formatting for source files and updates the prologs for files to include the name of the file being licensed. Signed-off-by: Darryl L. Pierce <dpierce at redhat.com> --- .gitignore | 12 +++++++++++ AUTHORS | 3 +- ChangeLog | 1 - INSTALL | 1 - Makefile.am

There was no valid reason to keep this set of APIs in their own agent and class, so those have been eliminated. A new HostListener method, HostListener::updated(), was added. It is via this method that the Host will notify all transports that the load average statistic is updated on the host. Signed-off-by: Darryl L. Pierce <dpierce at redhat.com> --- src/Makefile.am | 11

This is the first step towards separating the APIs from the means by which they are carried between the remote and local system. Added a few new classes and types: * HostTransport - defines a type for talking to the Host * Host - the public contract for the host APIs * HostAgent - the QMF transport layer Signed-off-by: Darryl L. Pierce <dpierce at redhat.com> --- .gitignore

The Processors agent reports details regarding the processing power for the managed node. Defined the agent within schema.xml. Wrote the wrapper code to create and populate an instance of the agent. The properties for the Processors agent have also been trimmed back to what is readily available via HAL. Signed-off-by: Darryl L. Pierce <dpierce at redhat.com> --- src/Makefile.am | 26

Hi all, I looked at patching oVirt Server to work with the latest Matahari schema, from Matahari's 'next' branch, version 0.4.0. ovirt-server depends on matahari version 0.0.5, which is currently latest in Fedora, and it is used during Node registration, to enumerate hardware information. Usage extracted from src/host-browser/host-register.rb is as follows: host_list =

Hey Y'all, I just installed the updates recommended by Software Update. Apr 25 19:04:43 Updated: xulrunner-10.0.4-1.el6.centos.i686 Apr 25 19:04:53 Updated: firefox-10.0.4-1.el6.centos.i686 Apr 25 19:05:03 Updated: thunderbird-10.0.4-1.el6.centos.i686 Apr 25 19:05:05 Updated: matahari-lib-0.4.4-12.el6_2.i686 Apr 25 19:05:06 Updated: matahari-agent-lib-0.4.4-12.el6_2.i686 Apr 25 19:05:09

It provides an initial implementation, LinuxPlatform, which is by default included in the build. Later code will refactor the build process to include the appropriate implementation depending on the target platform. Signed-off-by: Darryl L. Pierce <dpierce at redhat.com> --- src/Makefile.am | 4 ++ src/linux_platform.cpp | 117 ++++++++++++++++++++++++++++++++++++++++++++++++

The purpose of this patch is to add the matahari qmf agent package (http://git.et.redhat.com/?p=matahari.git;a=summary) to the node image. Matahari is a qmf agent that publishes hardware information to an amqp broker, replacing functionality from ovirt-identify-node. However, since the capability is not ovirt specific, the agent is maintained as a separate project (see link above). Dependencies

This patchset removes ovirt-identify-node and ovirt-listen awake. It adds startup script support for the matahari qmf agent, which takes over the responsibility for communicating node hardware capabilities to the ovirt-server. On the server side, host-browser has had its node identification functionality replaced by a new script, host-register, which is what interfaces with the matahari qmf agent

Just a quick script modeled after the libvirt-qpid example that dumps all connected ovirt-node hardware info to standard out, for debug purposes. --- src/qmf-matahari.example.rb | 50 +++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 50 insertions(+), 0 deletions(-) create mode 100644 src/qmf-matahari.example.rb diff --git a/src/qmf-matahari.example.rb

I meant to send this out a while back, but don't see that I ever actually sent it. I'm currently reworking the matahari agents to clean up the code, refactor elements and also enhance the agents. And the one that has me stumped is the NIC agent. The main problem is that, if we're going to represent network interfaces, we shouldn't limit it to only physical devices. Instead, I

i fired up a vm i haven't used in ages, it was centos 6.1, so I ran yum update -y, and it bombed on package conflicts with matahari and qpid-something... I had no need for matahari (heck, I don't even know what it is), so I removed matahari-\* and the yum update succeeded, but shouldn't yum update be able to update ANY centos 6 to latest ? this system has NO extra repos,