Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Hi again, On 02/01/2012 10:49 AM, Lawrence Strydom wrote: > woow ohkay. Thanks Felix. > > So that means I will have to use pacemaker or something similar to > automate the promotion if the primary fails? this is definitely recommended. You can always forego automatic failover. DRBD still gives you the ability to do a largely painless manual switch to node #2 (e.g. disaster recovery). > It almost seems easier to > set up dual primary. This doesn't really compute. On node failure, you will need to fail *something* over to the other node (even if that's only an IP address). What would you gain from dual-primary here? Mounting an ext2/3/4 partition on a second node will not work. OCFS2 or GFS would be options, but that still won't start your services on the second node on failure of the first. Finally, you need something to take care of fencing in your cluster when you go for dual-primary. So you will probably want pacemaker in any case, so you don't really save much effort. HTH, Felix PS. What problem are you trying to solve / what's your scenario? People will give you good advice on what type of setup would be most suitable for your needs if you share this info.