[ovs-discuss] 答复: ????: ????: [ovs]deleting interfaces connnected to ports results in ovs-vswitchd memory growing

pengyi Peng(Yi) pengyi.pengyi at huawei.com
Tue Apr 23 01:56:17 UTC 2013


I find that when ovs-vswitchd memory is abnormally growing, backlog in struct rpc is growing fast too. Almost every time in jsonrpc_run, stream_send returns -EAGAIN, the msg holds and waits next time to send. The backlog pluses the strlen of next msg. So the backlog can be as big as 1228252. But after all ports are deleted, the backlog will slowly decrease and in the end be zero again. However, at this time, with top command, the memory of ovs-vswitchd doesn't change at all. It is still bigger than the memory before all interfaces are deleted.

And last night, I tested another situation. 
First I create 200 ports which try to connect to interfaces that don't exist.
(ovs-vsctl --timeout=5 --no-wait add-port br-0 tap-no-exist$i vlan_mode=access tag=0 -- set interface tap-no-exist$i type=system)

Second, I create 1000 ports connected to interfaces that exist.
( tunctl -t vif$i
ifconfig vif$i up
ovs-vsctl --timeout=5 --no-wait add-port br-0 vif$i vlan_mode=access tag=0 -- set interface vif$i type=system)

After all ports finish connecting to interfaces, the memory of ovs-vswitchd is 530832KB. Today it changes to 524444KB. But the backlog in struct rpc is still very big(455532592). 

So I have two more questions. When will the ovs-vswitchd deal with the backlog successfully?? Is top command accurate to reflect the memory changes??

-----邮件原件-----
发件人: Ben Pfaff [mailto:blp at nicira.com] 
发送时间: 2013年4月19日 23:53
收件人: pengyi Peng(Yi)
抄送: discuss at openvswitch.org; Liuyongan; Zhangkun (K); dev at openvswitch.org
主题: Re: ????: ????: [ovs-discuss] [ovs]deleting interfaces connnected to ports results in ovs-vswitchd memory growing

On Fri, Apr 19, 2013 at 01:40:14AM +0000, pengyi Peng(Yi) wrote:
> I have attached the output of ovs-parse-leaks in the first
> E-mail. Here I attach it again.

Thanks.  I'll have a look a little later.


More information about the discuss mailing list