[openstack-dev] Propose BP for Common Heartbeat Driver

Xu He Jie xuhj at linux.vnet.ibm.com
Tue Jun 4 08:17:49 UTC 2013


On 2013?06?03? 00:41, Alex Glikson wrote:
> If the goal is heartbeat of OpenStack services (such as nova-compute, 
> cinder-volume, etc) -- then servicegroup is designed to do exactly 
> that (for historical reasons the name is a bit more generic -- 
> implying that it can eventually support additional group communication 
> functions for services). Is there anything in particular you see 
> missing in it?
Yes, the goal is heartbeat of openstack services.
> BTW, a more generic heartbeat abstraction might be useful for 
> scenarios like VM HA, where in addition to monitoring individual 
> services it could make sense to monitor the liveness of physical nodes 
> in a more holistic way (e.g., think of Linux-HA heartbeat/coresync as 
> a backend driver) 
Does 'VM HA' means the physical nodes that running the VMs is down, then 
nova should startup those VMs on another physical nodes?

What I understand was there are two layers of HA:
1.The HA of physical node was done by heartbeat/corosync+pacemaker, as 
this describe 
http://docs.openstack.org/trunk/openstack-ha/content/ch-intro.html
2.The HA of service that provided by openstack was done by this common 
heartbeat driver(to monitoring individual services).

The VM(compute service) provided by nova. So VM HA should be done by 
this common heartbeat driver, it needn't monitor liveness of physical 
nodes in heartbeat driver.
Does make sense?

I agree with you the heartbeat abstraction should be generic, then it 
can be used by more scenarios. I have written the document about the 
abstraction:
https://docs.google.com/document/d/1IClo_8OxG8b7bT9zcVdO5rIqucR6ZYq3yFhIj7UJIeA/edit#heading=h.2f5wiqa9s5j3
And there is a WIP patch: https://review.openstack.org/#/c/31048/

Welcome any comment!

Thanks.
>
> Regards,
> Alex
>
>
>
>
> From: Xu He Jie <xuhj at linux.vnet.ibm.com>
> To: OpenStack Development Mailing List 
> <openstack-dev at lists.openstack.org>,
> Date: 31/05/2013 03:46 PM
> Subject: [openstack-dev] Propose BP for Common Heartbeat Driver
> ------------------------------------------------------------------------
>
>
>
> Hi, guys,
>
> I created a blueprint to create common heartbeat driver that can used 
> by all the projects._
> __https://blueprints.launchpad.net/oslo/+spec/common-heartbeat-driver_
>
> The initial reason for creating this bp was quantum agent's heartbeat 
> based on DB. That is inefficient
> and not scalable.
>
> We already have solution for same reason in nova servicegroup api and 
> Matchmaker. So hope that can
> be merged as common module used by all the projects.
>
> The common heartbeat driver should concern several problem as below:
> * First... can be used by all the projects.
> * Driver interface can support existed db based heartbeat for bw 
> compatiblity.
> * Working with Matchmaker that used by zmq rpc backend. Matchmaker 
> will run heartbeat for each
> registered rpc consumer. Actually those heartbeat and agent/node 
> heartbeat run in same process.
> That is inefficient.
>
> Before the bp reviewed and approved, I also submited WIP code for 
> this. Hope that good for review bp
> and discussion. _https://review.openstack.org/#/c/31048/1_
>
> Thanks.
> Alex
>
> _______________________________________________
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
>
>
> _______________________________________________
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20130604/3f54e3fb/attachment.html>


More information about the OpenStack-dev mailing list