[Linux-HA] Groups vs colocations.... etc

[Andrew Beekhof]

On 11/27/06, Andre van der Vlies <andre at vandervlies.xs4all.nl> wrote:
>
> Andrew Beekhof wrote:
> > On 11/27/06, Andre van der Vlies <andre at vandervlies.xs4all.nl> wrote:
> >>
> >> I am setting up a 3 node cluster.
> >>
> >> the cluster has two IP-addresses (external and internal
> >> there are two NFS-filesystems..
> >> and there is postgres...
> >>
> >> All of this stuff has to run on one node.
> >>
> >> So I've put everything in a group...
> >>
> >> And found that the 'fail-count' is raised only on the last resource for
> >> that group.
> >>
> >> If one resource fails, all (following) resources are restarted (stopped
> >> &
> >> started) in sequence.
> >>
> >> Apparently postgres always fails if it is started; think I need to raise
> >> the timeout...
> >>
> >> There's no 'fail-count' for the group; accumulation of all is's
> >> resources...
> >>
> >>
> >> What I want is a stickiness for my resources; try 2 or 3 times to get a
> >> resource running.
> >
> >
> >> What don't want is a (running) resource to be stopped if another
> >> (higher)
> >> resource in that group fails...
> >>
> >> I want the group to fail(over) if one of it's resources runs out of its
> >> 'stickiness'...
> >
> > you cant have both
> >
>
> So, given:
>   IPaddr_1
>   IPaddr_2
>   NFS_1
>   NFS_2
>   PG
>
> there's no way I can prevent NFS_2 and PG from being stopped and started
> if NFS_1 fails, make NFS_1 retry 5 times and if it doesn't succeed the
> whole group needs to failover...  :-/


not in a group.
but groups are only a syntactic shortcut for a bunch of colocation and
ordering constraints.

so dont use a group and dont make NFS_2 depend on NFS_1