On 10/23/07, Riccardo Perni <[EMAIL PROTECTED]> wrote: > > Andrew Beekhof <[EMAIL PROTECTED]> ha scritto: > > > On 10/23/07, Riccardo Perni <[EMAIL PROTECTED]> wrote: > >> > >> > >> Andrew Beekhof <[EMAIL PROTECTED]> ha scritto: > >> > >> > On 10/22/07, Riccardo Perni <[EMAIL PROTECTED]> wrote: > >> >> >> Is it possible > >> >> >> to handle this situation? > >> >> > > >> >> > You may try quorumd. See > >> >> > > >> >> > http://www.linux-ha.org/QuorumServerGuide > >> >> > >> >> I'm going to look at it, but is'n it another SPOF? > >> > > >> > by definition, no. > >> > because you've already had at least one failure before quorumd > >> > becomes relevant > >> > >> Do you mean that the cluster will continue to work even if I have a > >> failure on the quorum server? > > > > my understanding is that the quorum server is not used unless you > > already dont have quorum... at which point you've lost half your nodes > > anyway > > Uhm, but at this point I already have a split-brain condition... or not?
Yes, and at this point each node aks the quorumd (running on an external third node) for quorum and only one node gets the ok. So one node starts all services and the second node without quorum stops all services. Regards, Andreas > > -- > Riccardo Perni > Unità Operativa Informatica Aziendale > ASL Roma-B > > > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
