> The object servers should only be talking to each other during replication.  
> They should not talk to the proxy, and probably not the load balancer.  Can 
> you provide the output of "swift-ring-builder /etc/swift/object.builder" and 
> more details on the network configuration of this system.

First please allow me to correct my first email. I found out that most of the 
traffic is REPLICATE requests and not rsyncs between storage nodes.

So here is the output of the ring builder:

# swift-ring-builder /etc/swift/object.builder 
/etc/swift/object.builder, build version 5
262144 partitions, 3.000000 replicas, 1 regions, 5 zones, 5 devices, 0.00 
balance
The minimum number of hours before a partition can be reassigned is 1
Devices:    id  region  zone      ip address  port  replication ip  replication 
port      name weight partitions balance meta
             0       1     5      10.10.6.16  6000      10.10.6.16              
6000   device4  10.00     157287    0.00 
             1       1     3      10.10.6.14  6000      10.10.6.14              
6000   device2  10.00     157286   -0.00 
             2       1     1      10.10.6.17  6000      10.10.6.17              
6000   device5  10.00     157286   -0.00 
             3       1     2      10.10.4.21  6000      10.10.4.21              
6000   device1  10.00     157287    0.00 
             4       1     4      10.10.6.15  6000      10.10.6.15              
6000   device3  10.00     157286   -0.00 

Well, as far as network configuration goes, I'm running Openstack on two 
blades. The second one has only a compute node and Swift installed (1 Proxy 
Node and 1 Storage Node). Both machines are on the same vlan. The others 
Storage Nodes are on a different vlan as guest virtual machines of the first 
two. Between the two vlans there's a firewall where I logged the traffic.

It came to my attention that this should be the object-replicator doing its 
work, but the fact is that I get about a thousand or 2 requests per second 
between vlans - (the real machine running a Proxy and a Storage Node) and the 
virtual machines (all Storage Nodes only) - and it consumes a lot of resources 
(mainly CPU) seems to be a bit of a waste because the cluster is not being used 
for more than 2 weeks - nothing added or deleted from it.

Can you tell me if this is normal behaviour? If so, how will this scale when I 
add more objects? Will it keep getting more and more CPU usage?

> Generally the proxies are configured with multiple interfaces as the object 
> servers and other backend storage nodes should not be accessible to clients 
> directly.

Yes, I believe this is the case.

Thank you for your answer

Diogo Vieira <[email protected]>
Programador 
Eurotux Informática, S.A. | www.eurotux.com
(t) +351 253 680 300
_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to     : [email protected]
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

Reply via email to