While we're at it, a cluster should generally have redundant cluster links (two physically separate network links with corresponding interface sections in the corosync configuration), especially if one of those links is also used as the DRBD replication link. Otherwise, any loss of communication on the single link that runs both, the cluster communication and DRBD replication, will typically cause a split brain situation. Additionally, fencing should be present to deal with any other situations where either a node fails or network communication fails. Fencing devices should be connected using an additional physically separate link, so that it still works if the cluster links do not.
Best regards, -- Robert ALTNOEDER - Software Developer [email protected] LINBIT | Keeping the Digital World Running DRBD HA - Disaster Recovery - Software-defined Storage DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. _______________________________________________ drbd-user mailing list [email protected] http://lists.linbit.com/mailman/listinfo/drbd-user
