Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Sun, 2007-02-11 at 18:50 -0600, Dan Brown wrote: > So I'm in the process of upgrading my DRBD setup from DRBD 0.7 to DRBD 8.0 > so that I can use GFS and two primaries on a two node cluster in order to > allow two servers to share the same filesystem without having to serve them > out via NFS and complicate things by having the servers mount themselves via > NFS. In my experience, mounting the NFS filesystem back on itself does increase the likelihood of problems during failover, since if your DRBD fails, then NFS fails, and the machine hangs (eventually). We use the bind option to mount the filesystem at the same path as the NFS path on the server that exports the NFS drive, so the filesystem layout matches all the other servers that do NFS mounts, but it decouples the NFS client during a failover event. <snip> > > And then a: > pvcreate /dev/drbd0 > vgcreate <fsgroup> > lvcreate -L <size> /dev/drbd0 > > However when I attempt the pvcreate I get the following error: > Device /dev/drbd0 not found (or ignored by filtering) > > How do I get around this? (or was my first method really the preferred one?) > Does the output of pvdisplay offer any hints? -- Jon