[ovs-dev] 答复: Re: [PATCH] The dependency between ovndb_servers-master andVirtualIP is wrong

[Numan Siddique]

Hi xurong,

Please see below for some comments.



On Wed, Jan 24, 2018 at 8:36 AM, <xu.rong at zte.com.cn> wrote:

> Hello,Liguoshuai:
>
> In my cluster environment, HA manages a lot of services, and VIP is not
> just for ovsdb.
>
> When I move ovsdb out of HA, VIP can't start, and other services are so
> abnormal, such as neutron-server.
>
>
>
>
> 原始邮件
> *发件人：* <ligs at dtdream.com>;
> *收件人：*赵静静00067370;徐熔00037997;
> *抄送人：* <russell at ovn.org>; <blp at ovn.org>; <azhou at ovn.org>; <
> dev at openvswitch.org>; <nusiddiq at redhat.com>;
> *日 期 ：*2018年01月24日 10:41
> *主 题 ：**Re: [ovs-dev] [PATCH] The dependency between ovndb_servers-master
> andVirtualIP is wrong*
> Hello, xurong, zhaojingjing:
>
> At that time, in order to avoid the problem of losing after the
> active/slave switchover,
> the slave node was promote prior to the start of the VIP.
> https://github.com/openvswitch/ovs/commit/a38f532320936147e6831153828af5
> bcba2fba54#diff-fb92408f16ddd415d9906e3976e60f47L255
>
> After that, the problem of data loss was completely solved in ovsdb,
> https://github.com/openvswitch/ovs/commit/05ac209a5d3a5e85896f58d16b244e
> 6b2a4cf2d0#diff-6878621809e8ea9210d85a69a5e06367
>
> I think it is possible to change this sequence again.
>
> Are there any new problems detected?
>
> >>> From: zhaojingjing0067370 <zhao.jingjing2 at zte.com.cn>
> >>>
> >>> ---
> >>>   Documentation/topics/integration.rst | 4 ++--
> >>>   1 file changed, 2 insertions(+), 2 deletions(-)
> >>>
> >>> diff --git a/Documentation/topics/integration.rst b/
> Documentation/topics/integration.rst
> >>> index 0447faf..d129e21 100644
> >>> --- a/Documentation/topics/integration.rst
> >>> +++ b/Documentation/topics/integration.rst
> >>> @@ -255,6 +255,6 @@ with the active server::
> >>>
> >>>       $ pcs resource create VirtualIP ocf:
> heartbeat:IPaddr2 ip=x.x.x.x \
> >>>           op monitor interval=30s
> >>> -    $ pcs constraint order promote ovndb_servers-
> master then VirtualIP
> >>> -    $ pcs constraint colocation add VirtualIP with
> master ovndb_servers-master \
> >>> +    $ pcs constraint order start VirtualIP then
> promote ovndb_servers-master
>

I think this makes sense.



> >>> +    $ pcs constraint colocation add master ovndb_
> servers-master with VirtualIP \
> >>>           score=INFINITY
>

In OpenStack tripleo installations, colocation constraint is set as
"VirtualIP with ovndb_servers-master".
Are you seeing issues with the below commands ?

 $ pcs constraint order start VirtualIP then promote ovndb_servers-master
 $ pcs constraint colocation add VirtualIP with master ovndb_servers-master

I would suggest to only change the "order" command and leave the colocation
command AS IS in your patch.

This is the pcs constraint show command in my tripleo environment.

****************
Ordering Constraints:
  start ip-192.168.24.12 then start haproxy-bundle (kind:Optional)
  start ip-10.0.0.6 then start haproxy-bundle (kind:Optional)
  start ip-172.16.2.4 then start haproxy-bundle (kind:Optional)
  start ip-172.16.2.10 then start haproxy-bundle (kind:Optional)
  start ip-172.16.1.5 then start haproxy-bundle (kind:Optional)
  start ip-172.16.3.11 then start haproxy-bundle (kind:Optional)
Colocation Constraints:
  ip-192.168.24.12 with haproxy-bundle (score:INFINITY)
  ip-10.0.0.6 with haproxy-bundle (score:INFINITY)
  ip-172.16.2.4 with haproxy-bundle (score:INFINITY)
  ip-172.16.2.10 with haproxy-bundle (score:INFINITY)
  ip-172.16.1.5 with haproxy-bundle (score:INFINITY)
  ip-172.16.3.11 with haproxy-bundle (score:INFINITY)
  ip-172.16.2.10 with ovn-dbs-bundle (score:INFINITY) (rsc-role:Started)
(with-rsc-role:Master)
Ticket Constraints:
**************************

Thanks
Numan


>>> --
> >>> 1.8.3.1
> >>>
> >>> _______________________________________________
> >>> dev mailing list
> >>> dev at openvswitch.org
> >>> https://mail.openvswitch.org/mailman/listinfo/ovs-dev
> >
> >
>
>
>