[Pacemaker] [ANNOUNCE] The Booth Cluster Ticket Manager - part of multi-site support in pacemaker

[Jiaju Zhang]

On Fri, 2012-02-17 at 17:54 +0900, 李泰勲 wrote:
> Hi, Jiaju
> 
> I am trying behavior of revoke Implementation manually.
> 
> I have questions of revoke.
> 
> If I revoked in one of grant ticket site, It would be free from grant
> status.
> but the other site runs automatic failover after that!
> Is it correct specifications?

Yes, currently the logic is as this. My original thought is to use the
ticket revoke command to do some maintenance work by the admin. One
possible scenario is the admin run the revoke command to revoke the
ticket from one site, and then do the maintenance work at that site.
If auto-failover has been configured, the ticket will failover to
another site to make the service always available. If auto-failover
has not been set, then ticket won't be failovered.

> 
> I think the expire timer is free in all of sites in that operation.
> What do you think about that?

I think what you think is reasonable, and current revoke logic might
confuse people;) So I'm thinking of changing it now.

> 
> https://bugzilla.novell.com/index.cgi
> I can't find anything of booth articles in that links.
> Is it right the links of the novell bugzilla?

Yes, that is the link for hosting bugs. But for the booth articles, it
is not there. a draft document for booth is at:

http://doc.opensuse.org/products/draft/SLE-HA/SLE-ha-guide_sd_draft/cha.ha.geo.html

And when the GEO cluster is released, there will be some revised
documents as well;)

Thanks,
Jiaju