On Thu, Sep 11, 2008 at 12:36:53PM +0200, randall wrote: > Christian Balzer wrote: > >Hello, > > > >On Thu, 11 Sep 2008 11:56:44 +0200 randall wrote: > > > >>what i would like to achieve is a fail over scenario with heartbeat-2 > >>and drbd8 on Etch > >> > >> > >I am doing exactly that (w/o LVM), using the latest versions from > >backports for HB/DRBD/vServer on a hand rolled kernel. > > > >See my mail in the archives here for a needed fix of the drbddisk > >script when used with vservers here: > >http://list.linux-vserver.org/archive?mss:835:200803:cgehldioambmojimggpf > > > >How you start up the vservers very much depends on your scenario, > >we wrote a (very simple and in need of enhancement before I would dare > >thrust it upon the world) HB resource agent that fires up all vservers > >under heartbeat control. We use a common big drbd resource and filesystem > >for all guests (and their configs), other people seem to prefer one > >resource per guest. > > > >Regards, > > > >Christian > > > thanks, > > this seems like something i could sync my teeth in. > the only difference indeed is the added logical volumes, i think i need > those for disk quota support from within one of the guests. that depends on the filesystem, and what you mean by disk quota support ... those which support disk limits on shared partitions will give you a per guest limit on the disk space used, but no user/group quota inside the guest, for that, you will need the separate partition you mentioned HTC, Herbert