[ovs-discuss] OVS Switch startup time

Sukhbir Singh sukhbir.in at gmail.com
Thu Apr 24 15:44:57 UTC 2014


Hi Ben,

* Ben Pfaff:

> On Thu, Apr 17, 2014 at 11:45:25AM -0400, Sukhbir Singh wrote:
>> I am resuming this thread from the earlier discussion about switch
>> startup time.  I decided to remove Mininet from the picture, just to
>> check how much time it takes for OVS to start a switch. To measure this,
>> I created a switch with:
>>
>> ovs-vsctl add-br s1 -- set Bridge s1
>> other_config:datapath-id=0000000000000001 -- set-fail-mode s1 secure --
>> add-port s1 s1-eth1 -- set Interface s1-eth1 ofport_request=1  --
>> add-port s1 s1-eth2 -- set Interface s1-eth2 ofport_request=2 --
>> set-controller s1 tcp:127.0.0.1:6633 ptcp:6634 -- set bridge s1
>> other-config:disable-in-band=true
>>
>> (Wrapped)
>>
>> I repeated this with the relevant changes to datapath_id, etc., for up
>> to 200 switches. Here are the start up times:
>>
>>         number of switches | time to start (seconds)
>>                        5   | .6
>>                        25  | 5
>>                        50  | 18
>>                        100 | 67
>>                        200 | 264
> 
> What did you actually measure?  The "ovs-vsctl" program run?  Something
> else?

Yes, basically: `time ./switches' where switches is just repeating the
above command for the number of switches in question (real time).

> How long does it take if you use "ovs-vsctl --no-wait"?

--no-wait does seem to help. The man pages says:

"Prevents  ovs-vsctl  from  waiting  for  ovs-vswitchd  to  reconfigure
 itself  according  to the the modified database."

This seems quite useful.

Also, is there a limit to the number of switches that can be created
assuming the usual resource constraints are in place?  (on the lines of
"Limits" in [0].)

[0] -
http://openvswitch.org/cgi-bin/ovsman.cgi?page=vswitchd%2Fovs-vswitchd.8

-- 
Sukhbir



More information about the discuss mailing list