On Friday, March 23, 2012 at 1:57 AM, ruslan usifov wrote: > I mean that, if throw pacemaker we organize fault tolerant monitor > (monitor that will work all time - even in fail case), we prevent > delays.
So the thing about this is that you can't use Pacemaker to start up different monitors in different places — monitors are required to persist their state and while you can add new nodes, you can't remove old ones. (This is required for the consistency and data safety guarantees that the monitors have to provide.) So you could use Pacemaker to restart crashed monitors, and you could (maybe? I'm not very familiar with it) use it to propagate information about which monitors you want new daemons to connect to, but you couldn't use it to start up new Monitors in a different location when old ones crash — Ceph will complain, because it has to. This shouldn't be a big problem for you in general — monitors are contacted on daemon startup but every other use of them is asynchronous and outside of data paths, so the only people that are really going to be impacted by a down monitor are administrators, who can handle it! -Greg -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
