[openstack-dev] [ceilometer] importutils problem in import_object_ns

Jiang, Yunhong yunhong.jiang at intel.com
Wed Nov 14 08:48:53 UTC 2012


> I uploaded a patch as https://review.openstack.org/#/c/15952/ for
> discussion/review. Should be "git review -D", but that command failed on my
> environment.
> 
> There are two issues on the rpc api method as described in the
> patch/comments. The nic/disk metadata possibly is available through
> extending nova-compute rpc_api, but for the instance object, I didn't find a
> clean solution.

I updated new patchset.
Thanks for Doug's patch for libvirt, which remind me that I don't need pass the real nova instance object, so I faked one. Considering we only depends on one interface from rpcapi, I think it's feasible for us to make sure this faked object is not broken.

So the issue left is the metadata. Followed is the information we can got from get_diagnostics. Although it provide all the static information that's presented by current libvirt pollster, it does not include any information for the virtual block/nic devices ,like MAC, IP, filter, DHCPSERVER, while is provided by current libvirt pollster. Also I didn't find method to get such information, other than call nova libvirt driver directly.

So my question is, should we create another rpcapi for this monitor specific purpose? I'd prefer to yes, since that will make life easier.

+------------------+------------+
| Property         | Value      |
+------------------+------------+
| cpu0_time        | 7580000000 |
| memory           | 524288     |
| vda_errors       | -1         |
| vda_read         | 262144     |
| vda_read_req     | 112        |
| vda_write        | 5606400    |
| vda_write_req    | 367        |
| vnet0_rx         | 1959353    |
| vnet0_rx_drop    | 0          |
| vnet0_rx_errors  | 0          |
| vnet0_rx_packets | 17673      |
| vnet0_tx         | 58983      |
| vnet0_tx_drop    | 0          |
| vnet0_tx_errors  | 0          |
| vnet0_tx_packets | 263        |
+------------------+------------
	

Thanks
--jyh





More information about the OpenStack-dev mailing list