AW: [LinuxFailSafe] drbd resource script patch

[Martin Bene]

> - FailSafe should not only control the primary, but also the=20
> secondary via a
>   resource.
>=20
>   (ie, the secondary is "online pending" until it is fully=20
> sync'ed etc)
>  =20
> - The primary resource _knows_ about this and can control the=20
> secondary
>   resource.
>=20
> - A special failover policy script will automatically pick=20
> the right node for
>   the primary.

While a good idea, this looks like way more work than I can currently =
put
into failsafe - esp. since the kludge used in the drbd resource scripts
actually works quite well :-) For more complex replicating systems this
aproach would probably be necessary. doing scripts for a replicating
database that works on more than 2 nodes could be fun :-)

There's another solution possible: the current failover scripts already =
have
"InPlaceRecovery" as one of the possible failover attributes for  =
failover.
just letting the exclusive script show "running on this node" while
replication is in progress and relying on "InPlace Recovery" to pick the
right node to start a resource group should work. I didb't use this =
because
default for recovery parameter is "Automatic Recovery", at least when =
using
the GUI to configure the cluster.=20

> A very minor nit pick - could you try to set your line-length=20
> in e-mail to 78 characters or so?

Argh, the blessings of MS exchange - sorry. (No I don't need any =
comments on
the combination of interest in high availability and the MUA/MTA =
combination
I'm using.) should be fixed now.

Bye, Martin