hmmm... @Arnold... you are correct. Thanks your reply "Yes, this can be made to work _if_ each mysql-instance has its own disk-space for the tables." this give me a hit...haha i think my mind is stuck at that time.
I already use active/Passive mode for my mysql cluster and it works fine now. thanks so much.... On Wed, Oct 9, 2013 at 3:56 AM, Arnold Krille <arn...@arnoldarts.de> wrote: > On Sun, 6 Oct 2013 02:08:24 +0800 Gray Wen Wen > <deutschland.g...@gmail.com> wrote: > > Hi all, > > now I am trying to configure a dual DRBD with mysql > > i wanna use active/active mode without any loadbalance. > > so my drbd is primary/primary on node1 and node2. > > the mount point is /mysql > > and i configure everything for mysql > > then when I start the pacemaker service > > the mysql can only be active on one node. > > and another server show mysql startup error. > > > > I tried few times...so here is my question, > > whether because of I set primay/primary drbd so actually the two > > nodes hold the same shared storage (/mysql folder on two nodes) > > so if one node start mysql that means it locks the folder /mysql > > and another node can not continue use this shared storage. > > > > does mysql active/active mode must use loadblance or something else? > > How do you think mysql (or any other database) works? > What happens when a row is added to a table? Now think about making > mysql sync that process (adding a row to a table) across several > instances, regardless whether these are all read-write masters or some > just read-only slaves. Yes, this can be made to work _if_ each > mysql-instance has its own disk-space for the tables. But these > mysql-instances will get horribly confused when someone else modifies > their disk-space without them knowing. Even if that "someone else" is > another mysql-instance meaning to do the right thing. There is a reason > why the second instance sees the tables as "locked". > > Long story short: Don't do what you are trying to do. > Either use drbd in active-passive and run _one_ mysql-instance on the > drbd-master. Or run one mysql instance with its own disk-space on each > node and promote/demote the write-master/read-only-slave with the > mysql-RA. > > Good luck, > > Arnold > > _______________________________________________ > Pacemaker mailing list: Pacemaker@oss.clusterlabs.org > http://oss.clusterlabs.org/mailman/listinfo/pacemaker > > Project Home: http://www.clusterlabs.org > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf > Bugs: http://bugs.clusterlabs.org > >
_______________________________________________ Pacemaker mailing list: Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
