Ovirt devel - May 2010 - Supercedes previous patch...

The previous patch, for whatever reason, fails to apply on top of tip.

This patch is reimplemented and, hopefully, fixed.

Signed-off-by: Darryl L. Pierce <dpierce at redhat.com>
---
 src/Makefile.am       |    2 ++
 src/host.cpp          |    9 ++++++---
 src/qmf/hostagent.cpp |    5 ++++-
 src/qmf/hostagent.h   |    7 +++++--
 src/schema.xml        |    6 ++++++
 5 files changed, 23 insertions(+), 6 deletions(-)

diff --git a/src/Makefile.am b/src/Makefile.am
index 498ee72..d830031 100644
--- a/src/Makefile.am
+++ b/src/Makefile.am
@@ -5,6 +5,8 @@ sbin_PROGRAMS = matahari
 first = qmf/com/redhat/matahari/Host.cpp
 
 generated_file_list = \
+	qmf/com/redhat/matahari/EventHeartbeat.cpp \
+	qmf/com/redhat/matahari/EventHeartbeat.h \
 	qmf/com/redhat/matahari/Host.h \
 	qmf/com/redhat/matahari/NetworkDevice.cpp \
 	qmf/com/redhat/matahari/NetworkDevice.h \
diff --git a/src/host.cpp b/src/host.cpp
index 29b84ec..d9bd698 100644
--- a/src/host.cpp
+++ b/src/host.cpp
@@ -101,11 +101,14 @@ Host::update()
 {
   _processors.update();
 
-  for(vector<NetworkDeviceAgent>::iterator iter =
_networkdevices.begin();
-      iter != _networkdevices.end();
+  time_t __time;
+  time(&__time);
+
+  for(set<HostListener*>::iterator iter = _listeners.begin();
+      iter != _listeners.end();
       iter++)
     {
-      iter->update();
+      (*iter)->heartbeat((unsigned long)__time);
     }
 }
 
diff --git a/src/qmf/hostagent.cpp b/src/qmf/hostagent.cpp
index 9807f65..7376f69 100644
--- a/src/qmf/hostagent.cpp
+++ b/src/qmf/hostagent.cpp
@@ -20,7 +20,7 @@
 #include "hostagent.h"
 #include <qpid/agent/ManagementAgent.h>
 
-namespace _qmf = qmf::com::redhat::matahari;
+#include "qmf/com/redhat/matahari/EventHeartbeat.h"
 
 HostAgent::HostAgent(Host& host)
   :_host(host)
@@ -35,6 +35,8 @@ HostAgent::~HostAgent()
 void
 HostAgent::setup(ManagementAgent* agent)
 {
+  this->_agent = agent;
+
   _management_object = new _qmf::Host(agent, this);
   agent->addObject(_management_object);
 
@@ -65,4 +67,5 @@ HostAgent::ManagementMethod(uint32_t method, Args&
arguments, string& text)
 void
 HostAgent::heartbeat(unsigned long timestamp)
 {
+  this->_agent->raiseEvent(_qmf::EventHeartbeat(timestamp));
 }
diff --git a/src/qmf/hostagent.h b/src/qmf/hostagent.h
index 6d2f2a0..f9a6f8e 100644
--- a/src/qmf/hostagent.h
+++ b/src/qmf/hostagent.h
@@ -31,11 +31,14 @@
 using namespace qpid::management;
 using namespace std;
 
+namespace _qmf = qmf::com::redhat::matahari;
+
 class HostAgent : public Manageable, public HostListener
 {
  private:
-  qmf::com::redhat::matahari::Host* _management_object;
-  Host& _host;
+  _qmf::Host*      _management_object;
+  Host&            _host;
+  ManagementAgent* _agent;
 
  public:
   HostAgent(Host& host);
diff --git a/src/schema.xml b/src/schema.xml
index 206e39b..31abb35 100644
--- a/src/schema.xml
+++ b/src/schema.xml
@@ -14,6 +14,12 @@
 
   </class>
 
+  <eventArguments>
+    <arg name="timestamp" type="absTime" />
+  </eventArguments>
+
+  <event name="heartbeat" args="timestamp" />
+
   <!-- The processor for the node. -->
   <class name="Processors">
     <property name="host"   type="objId"
access="RC" desc="The host machine." index="y"
references="Host" parentRef="y" />
-- 
1.6.6.1

On Wed, May 26, 2010 at 09:23:44AM +0200, Andrew Beekhof
wrote:
> Some quick questions...
> 
> Is the interval not being used yet?  I don't see any timers or such.
Not implemented yet. This patch just introduces the actual heartbeat event over
QMF.
> Wouldn't you still need to update the network devices in
Host::update()?
Those are going to get refactored. As it stands they weren't being updated.
> Does the heartbeat event need a counter (like ICMP packets do) to
> detect backward jumps and/or lost heartbeats?
Right now the heartbeat sends out the timestamp. But if the clock gets changed,
it
could seem to be a backwards jump. Changing it to be a counter instead would
be a simple change to this patch for pushing.

-- 
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/

On Thu, May 27, 2010 at 08:43:30AM +0200, Andrew Beekhof
wrote:
> >> Does the heartbeat event need a counter (like ICMP packets do) to
> >> detect backward jumps and/or lost heartbeats?
> > 
> > Right now the heartbeat sends out the timestamp. But if the clock gets
changed, it
> > could seem to be a backwards jump. Changing it to be a counter instead
would
> > be a simple change to this patch for pushing.
> 
> I think you'd probably want both.  The timestamp and a sequence number.
That's cool. Okay, so if this is an ACK with that change (I'll wait for
a response) I'll go ahead and push it.

-- 
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL:
<http://listman.redhat.com/archives/ovirt-devel/attachments/20100527/2417cf83/attachment.sig>