Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Monday 07 June 2010 08:41:42 Florian Haas wrote: > On 06/06/2010 12:18 PM, Olivier Le Cam wrote: > > Bart Coninckx wrote: > >> Sounds like a Heartbeat issue. Heartbeat should not make the node > >> primary before the sync has finished. > > That assumption is incorrect for this configuration. > > > OK, that confirm my though, thank you Bart. I probably missed to > > configure something! > > > > My haresources file only has this in: > > > > drbd1 IPaddr::192.168.19.125/24/eth1 \ > > drbddisk::r0 \ > > Filesystem::/dev/drbd0::/nfs3-exports::ext3::noatime \ > > killnfsd \ > > nfs-kernel-server > > Everything is working exactly as expected. Heartbeat in haresources > configuration with auto_failback on fails back as soon as possible and > will happily promote a DRBD device which is still being synchronized > (and whose local disk state is, thus, Inconsistent). > > Don't want that? Ditch your haresources setup and go for Pacemaker. > > Florian > Or have it configured in such a way that after a failover stays with the last node until you reboot that one. That was why I was asking for the ha.cf . I have it configured in such a way. It gives the other node also the chance to catch up. Didn't know this was default behaviour though ...