Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Usman Ahmad wrote: > The primary server waited for some time on boot up and then came up. > I want the auto-failback on ? Are you *really* sure? Are you 100% sure you know what it means? You should at least turn it off while you try to debug your setup. DRBD can work with it turned on; it just causes unnecessary transitions. Even if you have a "preferred" machine, it's probably best that you fail back to it manually once you're sure everything is fine. In a peer cluster there is rarely a need to automatically fail back to one machine as the whole point is they are peers. (That's not to say there aren't legitimate reasons to use it, just that there are many cases where it's not necessary) > Yes, the servers were connected, even if I shutdown the heratbeat the > disk is shifted to the secondary server, but they remain in standalone > mode, the drbd can talk to each other as the (Secondary server) changed > states from primary/unknown to secondary/unkown. This doesn't mean DRBD can communicate. On the contrary, the "Unknown" means that it can't. The fact it changed state is probably because Heartbeat changed the state of DRBD. Have you checked your firewall? It's common for people to forget to allow DRBD through the firewall between the nodes. > I do not understand why? can't find much in the logs, what thing should > I look for specifically? Anything to do with DRBD in the first instance! Just read the logs and follow the logical process through. DRBD/Heartbeat are both quite verbose and tells you lots about what they are doing. After you do that, if you still don't understand, point us towards the specific items in the log where you don't understand. > Is it ok to start the heartbeat and the drbd processes at boot time > automatically? Yes, and you probably want to do that. Tim